There was a fellow from MegaCorp (manager, not a programmer) who did the
conference to check out Lisp for a new $5m project. It's a C++ shop and
they are not in love with the overnight builds. One guy in his group has
been "playing with Lisp" and encouraged its use in the new project, so
this guy came to check us out. (Or get a week in Frisco on the company's
dime. <g>) Anyway, he was worried a lot about justifying Lisp to other
management, even if he concluded Lisp would be better.
One specific: he said the ALU should cook up a certification exam. He
understood that certification might be a joke in our domain, but that
nevertheless it was the kind of thing that would make Lisp look more
respectable to MegaCorp types.
He even seemed to think a simple timed on-line test would do the trick,
though how you stop cheating I do not know.
This was a new one on me, and was only one bloke's input. I scoff at
certification for programmers, but if it would be easy to implement and
if folks in tall buildings care (two big ifs) why not?
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
Kenny Tilton <·······@nyc.rr.com> wrote in news:3DC40755.3060803
@nyc.rr.com:
> There was a fellow from MegaCorp (manager, not a programmer) who did the
> conference to check out Lisp for a new $5m project. It's a C++ shop and
> they are not in love with the overnight builds. One guy in his group has
> been "playing with Lisp" and encouraged its use in the new project, so
> this guy came to check us out. (Or get a week in Frisco on the company's
> dime. <g>) Anyway, he was worried a lot about justifying Lisp to other
> management, even if he concluded Lisp would be better.
>
> One specific: he said the ALU should cook up a certification exam. He
> understood that certification might be a joke in our domain, but that
> nevertheless it was the kind of thing that would make Lisp look more
> respectable to MegaCorp types.
>
> He even seemed to think a simple timed on-line test would do the trick,
> though how you stop cheating I do not know.
>
> This was a new one on me, and was only one bloke's input. I scoff at
> certification for programmers, but if it would be easy to implement and
> if folks in tall buildings care (two big ifs) why not?
>
If it was done for real look at how www.giac.org does it. You need to
write a paper and pass before you get to sit for the test. Then your paper
is posted on the web so that people can review your work. It is a good
deal of work to get this done though. The brainbench type web tests are
things I would not put on my resumee.
marc
* Kenny Tilton
| One specific: he said the ALU should cook up a certification exam.
How about just asking people to give a paper at a Lisp conference and
making sure that the quality standards are sufficiently high?
http://www.cs.ucsb.edu/~almeroth/conf/stats/
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> wrote in ·····················@naggum.no:
> * Kenny Tilton
>| One specific: he said the ALU should cook up a certification exam.
>
> How about just asking people to give a paper at a Lisp conference and
> making sure that the quality standards are sufficiently high?
>
> http://www.cs.ucsb.edu/~almeroth/conf/stats/
>
I do not know, a certifacation is suposed to demonstrate a certian min
level of mastery consistantly. Your way would allow the people who make
the cut to put it down, much as they do today, but what of the people who's
papers were good enough to be used but not used due to other issues?
I am not saying it is not a good metric, I think it is. But that it is too
subject to outside influences for this purpose.
marc
* Marc Spitzer
| I am not saying it is not a good metric, I think it is. But that it is
| too subject to outside influences for this purpose.
I must admit to some ulterior motives. First, if this was the requirement,
I would get off the hook because I have given a paper at a Lisp conference.
But (more) importantly, it would make a lot of people submit papers to Lisp
conferences and thus would make the conferences more interesting and more
frequent. Fundamentally, I do not see the point with certification if it
is a "selfish" measure, i.e., one where the community benefit of having one
more certified programmer is negative, which it would be if the purpose was
to make it easier for managers to replace one Common Lisp programmer with
another or have more people compete for the same jobs. If managers want
that, they can have it, but giving it to them should benefit the community
more than the particular manager. Otherwise, managers get a strangle-hold
on the market and will work hard to /lower/ the certification requirements
so that they have more people to choose from and can lower their costs.
The entrance fee to the Certified Common Lisp Programmer market should be
that you have done something that clearly benefits the existing certified
programmers, not just something that benefits the /future/ candidates. The
same rationale underlies the requirements to grant professional degrees.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> wrote in ·····················@naggum.no:
> * Marc Spitzer
>| I am not saying it is not a good metric, I think it is. But that it
>| is too subject to outside influences for this purpose.
>
> I must admit to some ulterior motives. First, if this was the
> requirement, I would get off the hook because I have given a paper
> at a Lisp conference. But (more) importantly, it would make a lot of
> people submit papers to Lisp conferences and thus would make the
> conferences more interesting and more frequent.
Yes that would be good. And being grandfathered is always nice.
> Fundamentally, I do not see the point with certification if it is a
> "selfish" measure, i.e., one where the community benefit of having
> one more certified programmer is negative, which it would be if the
> purpose was to make it easier for managers to replace one Common
> Lisp programmer with another or have more people compete for the
> same jobs. If managers want that, they can have it, but giving it
> to them should benefit the community more than the particular
> manager. Otherwise, managers get a strangle-hold on the market and
> will work hard to /lower/ the certification requirements so that
> they have more people to choose from and can lower their costs. The
> entrance fee to the Certified Common Lisp Programmer market should
> be that you have done something that clearly benefits the existing
> certified programmers, not just something that benefits the /future/
> candidates. The same rationale underlies the requirements to grant
> professional degrees.
>
Those are good points that I had not considdered, thanks for bringing
them up.
What would you think of a tiered approach, apprentice -> journeyman ->
master setup. Where the journeyman and master grades would have the
requirement to help current advanced members. I believe I am talking
about a guild or guildish thing.
Also there are certian certs out there that having makes you more money,
CCIE comes to mind. It is a very hard test, here is how it works:
1: take a written test
2: iff you pass then you go to cisco and:
a: day 1 build a network
b: if you built it right then day 2 fix it after they break it.
It is designed to fail people.
I guess the trick is not to let the managers have much, if any, say in
the matter.
marc
Marc Spitzer <········@optonline.net> writes:
> Also there are certian certs out there that having makes you more
> money, CCIE comes to mind. It is a very hard test, here is how it
> works:
>
> 1: take a written test
This is known as the Qualifying Exam and is a 2 hour
computer-mediated multiple-choice exam.
> 2: iff you pass then you go to cisco and:
> a: day 1 build a network
> b: if you built it right then day 2 fix it after they break it.
This has changed. It is no longer a 2 day lab exam - it is just 1
day. Furthermore, the troubleshooting section has been removed.
> It is designed to fail people.
This has not changed. The average CCIE fails the lab 2 times and
passes on the 3rd time.
There were some initial indications that switching from a 2 day
format to a 1 day format actually made the lab *harder* judged by
the # of attempts/# of passes per day.
> I guess the trick is not to let the managers have much, if any, say
> in the matter.
Even the CCIE is not a great example of a certification "worth
having". :-) For starters, Cisco created it to assist management
in substituting one network engineer for another, without requiring
management to be clue-full enough to tell how good an engineer is.
This motivation provided a powerful incentive to Cisco to make it
as easy as possible [without losing all credibility] for people to
begin the CCIE certification process. The end result: a flood of
(relatively) un-experienced individuals who are certified as
"experts" because they learned a lot about how to work with Cisco
networking products.
For example, one minimum requirement to take the lab is 2 years of
full-time professional experience. Many CCIEs have only 3 or 4
years of experience. Now, with the downturn in the economy, there
is a glut of CCIEs - making the cert worth less for those of us who
hold it.
Moral: don't let a vendor control the certification either...
-jon
--
------------------
Jon Allen Boone
········@delamancha.org
CCIE #8338
Erik Naggum wrote:
> * Kenny Tilton
> | One specific: he said the ALU should cook up a certification exam.
>
> How about just asking people to give a paper at a Lisp conference and
> making sure that the quality standards are sufficiently high?
If we are doing a sick form of the Turing Test in which Lisp tries to
fool suits into thinking we are Java, then our certfication process must
look like theirs. I know one criterion is that it must be something
Asians can pass by memorizing a book, so the original content required
of a paper might give us away as Not Java.
Is there a metacertification exam template? Do folks write code during
the exam (as in more than just a line or two)? It would be great fun
here on cll to brainstorm an exam if we knew what it should look like.
btw, the gent from the big company (i think it was a phone company) said
the certification thing itself might draw folks towards CL. A new exam
announced on SlashDot and to the IT press in general in and of itself
would make clear that the news of CL's death (in Wired?) was greatly
exagerated.
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
On Sat, 2 Nov 2002, Kenny Tilton wrote:
> If we are doing a sick form of the Turing Test in which Lisp tries to
> fool suits into thinking we are Java, then our certfication process must
> look like theirs.
I hate to advertise my age, but once upon a time, that "sick form of the
Turing Test" was called a "computer science degree". Doesn't that suffice
for certification anymore? Or does no university teach lisp anymore?
Or does noone get computer science degrees anymore?
> I know one criterion is that it must be something
> Asians can pass by memorizing a book, so the original content required
> of a paper might give us away as Not Java.
Whether or not this insult came from an Asian as you have said elsewhere,
other Asians reading this newsgroup don't know that, and do indeed find
it insulting. Why not err on the side of politeness? Or does that violate
your right to be a moron? Oh gosh, I've gone and said something
politically incorrect again.....
> kenny tilton
> clinisys, inc
* Nick Geovanis
| IT Computing Svcs Computing's central challenge:
| Northwestern Univ How not to make a mess of it.
| ··········@nwu.edu -- Edsger Dijkstra
+------------------->
Nicholas Geovanis <·······@merle.acns.nwu.edu> wrote in
·············································@merle.acns.nwu.edu:
> On Sat, 2 Nov 2002, Kenny Tilton wrote:
>
>> If we are doing a sick form of the Turing Test in which Lisp tries to
>> fool suits into thinking we are Java, then our certfication process
>> must look like theirs.
>
> I hate to advertise my age, but once upon a time, that "sick form of
> the Turing Test" was called a "computer science degree". Doesn't that
> suffice for certification anymore? Or does no university teach lisp
> anymore? Or does noone get computer science degrees anymore?
It used to be you hired a person and you trained them in what you needed
that they did not know, within certian limits. The reason why this made
sence is that you would keep this person for 5-20+ years. Now the
average IT job lasts around 2 years in my part of the US at least. This
means that dumb managers hire employees by skills not as people. Java
certification get you through the first pass of resume screening, the OCR
phase.
>
>> I know one criterion is that it must be something
>> Asians can pass by memorizing a book, so the original content
>> required of a paper might give us away as Not Java.
>
> Whether or not this insult came from an Asian as you have said
> elsewhere, other Asians reading this newsgroup don't know that, and do
> indeed find it insulting. Why not err on the side of politeness? Or
> does that violate your right to be a moron? Oh gosh, I've gone and
> said something politically incorrect again.....
I think the memorize thing comes from 2 things that I am cable TV(history
channel, the learning channel and discover type education) knoledgable
about. The first is the work needed to be literate in Chinese, you must
memorize each word. And second a some what different view on knowledge,
different views were allowed to coexist and were recorded and referenced.
Also for a very long time the Chinese civil service test were based on
memorization and applacation of certian classics, Confusious and others I
dont know about. It was like passing the bar, but the test lasted 3
straight days.
marc
>> kenny tilton
>> clinisys, inc
>
> * Nick Geovanis
>| IT Computing Svcs Computing's central challenge:
>| Northwestern Univ How not to make a mess of it.
>| ··········@nwu.edu -- Edsger Dijkstra
> +------------------->
>
>
Nicholas Geovanis wrote:
> On Sat, 2 Nov 2002, Kenny Tilton wrote:
>
>
>>If we are doing a sick form of the Turing Test in which Lisp tries to
>>fool suits into thinking we are Java, then our certfication process must
>>look like theirs.
>
>
> I hate to advertise my age, but once upon a time, that "sick form of the
> Turing Test" was called a "computer science degree". Doesn't that suffice
> for certification anymore?
I think the key was that it be language-specific, so a degree by itself
would not serve. Lisp on the transcript? Well, do employers get
transcripts? And from what I hear, in other professions it is one thing
to get the degree, quite another to pass the professional certification
exams, so mebbe the certification is a higher standard?
>
>>I know one criterion is that it must be something
>>Asians can pass by memorizing a book, so the original content required
>>of a paper might give us away as Not Java.
>
>
> Whether or not this insult came from an Asian as you have said elsewhere,
> other Asians reading this newsgroup don't know that, and do indeed find
> it insulting.
And they all asked you to tell me so? Speaking of which, the first
person to whine about my remark suggested the "Asians" he knew did /not/
have excellent study habits. Right. Some folks don't like immigration, I
think it's the only thing keeping the US competitive; I sure wouldn't
want to count on American students to keep us ahead intellectually.
I heard once that some top school was placing inverse quotas on asian
enrollment, because us honkies could not get in on merit. I say, screw us.
btw, it is hard to imagine study habits being a racial outcome. Clearly
there is simply tremendous scholastic competitiveness in many Asian
countries and kids just tend to end up with that work ethic.
It's like Philadelphia fighters. Why are they so tough? Because in
Philadelphia gyms an ethic has developed in which sparring is as intense
as a title fight.
I can see how some might think I was saying that the /only/ thing Asians
bring to the table is a work ethic, but in truth I was just harking back
to my Asian friend's laughing, "We just memorize the book!"
You questioned the existence of my alleged friend. Does it matter? Is it
OK if the observation behind the joke came from an Asian? It's true that
when it comes to racial/ethnic/religious stuff there are jokes the group
targeted by the joke can make which others cannot, and I knew I was
flirting with that limitation when I made my remark.
To a degree my larger agenda is to stand up to humorless, knee-jerk PC
bores. It's a fine line between racism and , and maybe an NG is an
especially dangerous place to do that. I mean, some folks here know I am
rarely serious for more than a sentence or two in a row, but what about
newcomers?
What I finally decided was not to censor myself on behalf of the PCers,
rather to go ahead and make the quip and deal with the expected heat.
> Why not err on the side of politeness?
Where is Eminem when I need him?
> Or does that violate
> your right to be a moron? Oh gosh, I've gone and said something
> politically incorrect again.....
That's not politically incorrect, that's a personal characterization.
Show some balls, man, take on an entire ethnic group!
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
Kenny Tilton wrote:
>
> To a degree my larger agenda is to stand up to humorless, knee-jerk PC
> bores. It's a fine line between racism and [oops. insert "Don-Ricklesian pan-ethnic humor"], and maybe an NG is an
> especially dangerous place to do that.
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
Kenny Tilton <·······@nyc.rr.com> writes:
> Nicholas Geovanis wrote:
> > Or does that violate your right to be a moron? Oh gosh, I've gone
> > and said something politically incorrect again.....
>
> That's not politically incorrect, that's a personal
> characterization. Show some balls, man, take on an entire ethnic
> group!
According to the mentally challenged PC adepts you have to say
``mentally challenged�� AFAIR.
Regards,
--
Nils G�sche
"Don't ask for whom the <CTRL-G> tolls."
PGP key ID 0x0655CFA0
Nils Goesche wrote:
> Kenny Tilton <·······@nyc.rr.com> writes:
>
>
>>Nicholas Geovanis wrote:
>
>
>>>Or does that violate your right to be a moron? Oh gosh, I've gone
>>>and said something politically incorrect again.....
>>
>>That's not politically incorrect, that's a personal
>>characterization. Show some balls, man, take on an entire ethnic
>>group!
>
>
> According to the mentally challenged PC adepts you have to say
> ``mentally challenged�� AFAIR.
Well, I was playing dumb. Moron, idiot, and spastic are all technical
terms for folks with various handicaps (oops), so mebbe Nicholas /was/
taking on a group.
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
Kenny Tilton wrote:
>
> > Why not err on the side of politeness?
>
> Where is Eminem when I need him?
I think he's sorry he hurt his mama (she slipped on one of his toys) so
tonight he's cleaning up his bedroom.
* Kenny Tilton
| btw, it is hard to imagine study habits being a racial outcome. Clearly
| there is simply tremendous scholastic competitiveness in many Asian
| countries and kids just tend to end up with that work ethic.
There is also the parents' attitude. In some study I read a while ago,
so I cannot recall any reference to it, but I remember the details well,
80% of American parents thought their children did better than average at
school, while 10% of Japanese parents thought the same. Parents were
also asked whether they thought their children could get better grades if
they worked harder or whether the school/teacher could do something to
improve their children's grades. 75% of the American parents thought the
school/teacher could do a better job to "motivate" their children. 95%
of the Japanese parents thought their children could get better grades if
they worked harder. Asked whether they thought they could help their
children get better grades by helping them with their homework, 20% of
the American parents said they did help their children with homework, and
95% of the Japanese parents said they did. If you ask me, there is only
one cause for the dismal performance of the average American student:
Their parents. (Norway is no different, and probably worse. Over here,
virtually nobody believes their own life to be their own responsibility,
but instead believe that all the world's problems can be solved by pumping
money up from our oil-rich sea shore.)
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> writes:
> If you ask me, there is only one cause for the dismal performance
> of the average American student: Their parents.
Your're completely right for the German children as well. I am father
of two kids, the older one nine years old. He's good at school, but I
am sure he would have performed rather bad if we would not have
invested quite a lot of time and money in his education.
> (Norway is no different, and probably worse. Over here, virtually
> nobody believes their own life to be their own responsibility, but
> instead believe that all the world's problems can be solved by
> pumping money up from our oil-rich sea shore.)
You should have joined the EU. I guess that your money problems would
have been markedly reduced:-)
Nicolas.
* Erik Naggum wrote:
> How about just asking people to give a paper at a Lisp conference and
> making sure that the quality standards are sufficiently high?
I think the problem with this is that the skills needed to write
conference papers (typically dealing with some topic which can be
described well in a small number of pages read by a human - so little
confusions and vageuenesses are OK) are different than the skills
needed to write significant programs (typically dealing with a topic
which needs a much larger number of pages to describe, and which is
read by a machine, so must have no little confusions and vageuenesses
at all; however this machine readable description must *also* be
comprehensible to human readers!).
--tim
From: Christopher C. Stacy
Subject: Re: Conference moment: Lisp certification?
Date:
Message-ID: <u3cqiwser.fsf@dtpq.com>
Fortune 500 companies were one of the primary users of Lisp,
back when Lisp (and AI) was popular. Also, I am highly skeptical
that certification is what's keeping Lisp from being used there now.
I haven't talked with any such companies lately, but one data point
does not make the case for me.
Christopher C. Stacy wrote:
> Fortune 500 companies were one of the primary users of Lisp,
> back when Lisp (and AI) was popular. Also, I am highly skeptical
> that certification is what's keeping Lisp from being used there now.
Right, I think it was understood to be a small matter we might want to
consider, just another brick in the wall. The key is, if it is a
no-brainer and we are looking for something to improve CL's adoption,
why not? Give 'em what they want.
As it stands, CL gets laughed out of the running in IT, because the
immune system identifies us as "not-ready-for-IT". Maybe if we adorn
ouselves with a few ITish proteins like certification the immune system
will think we are friend, not foe.
And yes, it was just one opinion from one IT exec who had come to the
Lisp conference ostensibly because one of his people was arguing for
Lisp on the next project. Come to think of it, if we won't listen to
him, who should we listen to, to get into F500? It would be nice to get
more suits into this thread, or at least input from folks here who have
worked in tall buildings.
I have, and I can't say I heard anyone talking about certification.
otoh, i did know one IT recruiter who thought some guy was a genius
because he had passed a wadge of MS exams in short order (an Asian!).
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
"Kenny Tilton" <·······@nyc.rr.com> wrote in message ···················@nyc.rr.com...
>
>
> Christopher C. Stacy wrote:
> > Fortune 500 companies were one of the primary users of Lisp,
> > back when Lisp (and AI) was popular. Also, I am highly skeptical
> > that certification is what's keeping Lisp from being used there now.
>
> Right, I think it was understood to be a small matter we might want to
> consider, just another brick in the wall. The key is, if it is a
> no-brainer and we are looking for something to improve CL's adoption,
> why not? Give 'em what they want.
>
And all I'm saying is, do we really want to work for the forces of darkness?
Any company that has the expertise to determine their true needs and assess
programmer competence doesn't need these tests. Use of these tests is
an admission of managerial incompetence. Who wants to work for such a company?
I might do so only out of sheer desperation.
Is it possible that companies use these tests because they just happen to measure
exactly the skills that the company needs? No. I have taken many tests
in my life, and done quite well on most all of them, and I have never seen one that
measures much of anything other than the ability to do well on that particular test.
I know that this is a trite observance, but that doesn't make it any less true.
Programmer certification tests engender a situation akin to that created by "school
accountability" in the US. In an attempt to lift everyone up to a minimum standard of
competence, we tie funding to test results, and end up churning out droves of students that
have been taught nothing but how to pass the tests. It sucks for everyone in the end.
Don't cave in to immorality out of expediency.
-James Russell
A long time ago, in a galaxy far, far away, "J L Russell" <·········@alum.mit.edu> wrote:
> And all I'm saying is, do we really want to work for the forces of
> darkness?
>
> Any company that has the expertise to determine their true needs and
> assess programmer competence doesn't need these tests. Use of these
> tests is an admission of managerial incompetence. Who wants to work
> for such a company? I might do so only out of sheer desperation.
I'd be game to say this in a somewhat "less hostile" manner.
> Is it possible that companies use these tests because they just
> happen to measure exactly the skills that the company needs? No. I
> have taken many tests in my life, and done quite well on most all of
> them, and I have never seen one that measures much of anything other
> than the ability to do well on that particular test. I know that
> this is a trite observance, but that doesn't make it any less true.
The /useful/ tests I have taken have been administrated by someone
competent that was looking to see if I understood things.
The latest example was actually a couple of weeks ago, relating to
starting a contract. I "punted" on a couple questions, hacking a
timestamp into UTC because I never can remember without consulting
manuals how to do date computations in SQL. What was particularly
interesting was that the fellow reviewing the results commented that
some of the answers were unconventional, but that people usually got
the questions downright /wrong/.
And there probably lies the other piece: A /good/ test is one where
you'll get some answers wrong, because it has some questions tough
enough to challenge everyone. Which may be nicely contrasted with the
"certification" thing where 'it isn't a good certification unless
typical Asian "memorize 'til you drop" techniques can be used.'
> Don't cave in to immorality out of expediency.
.. But remember that by not being thusly expedient, you'll be largely
excluded from F500 environs. That's not necessarily a /bad/ thing,
but it's something to keep in mind.
--
(reverse (concatenate 'string ·············@" "sirhc"))
http://cbbrowne.com/info/unix.html
Rules of the Evil Overlord #64. "I will see a competent psychiatrist
and get cured of all extremely unusual phobias and bizarre compulsive
habits which could prove to be a disadvantage."
<http://www.eviloverlord.com/>
* J L Russell
| And all I'm saying is, do we really want to work for the forces of darkness?
I think that should be a personal decision. It would be better for all
of us if we did not have to check with what "we" want before each of us
can make a personal decision, and consequently it would be nice if those
who do not want something at least do not block the way for those who do.
| I have taken many tests in my life, and done quite well on most all of
| them, and I have never seen one that measures much of anything other than
| the ability to do well on that particular test. I know that this is a
| trite observance, but that doesn't make it any less true.
Tests are not useful for what they measure, but for how what they measure
correlates with other things. If some foot size happened to correlate
well with programmer proficiency, one could measure foot size and get
high programmer proficiency for no better reason than that other people
with the same foot size had high proficiency as programmers; one would
not measure proficiency as such. Even if foot size correlated weakly
with programmer proficiency, like 75% chance of getting a good programmer
with some foot size, it could still be more valuable than anything else
that had a lower correlation coefficient.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
"Erik Naggum" <····@naggum.no> wrote in message ·····················@naggum.no...
> * J L Russell
> | And all I'm saying is, do we really want to work for the forces of darkness?
>
> I think that should be a personal decision. It would be better for all
> of us if we did not have to check with what "we" want before each of us
> can make a personal decision, and consequently it would be nice if those
> who do not want something at least do not block the way for those who do.
>
Of course. I just phrased that poorly. Anyway, it's a rhetorical
exhortation, not a suggestion for prescriptive injunction.
> | I have taken many tests in my life, and done quite well on most all of
> | them, and I have never seen one that measures much of anything other than
> | the ability to do well on that particular test. I know that this is a
> | trite observance, but that doesn't make it any less true.
>
> Tests are not useful for what they measure, but for how what they measure
> correlates with other things. If some foot size happened to correlate
> well with programmer proficiency, one could measure foot size and get
> high programmer proficiency for no better reason than that other people
> with the same foot size had high proficiency as programmers; one would
> not measure proficiency as such. Even if foot size correlated weakly
> with programmer proficiency, like 75% chance of getting a good programmer
> with some foot size, it could still be more valuable than anything else
> that had a lower correlation coefficient.
>
Nor do I disagree with this. Of course there is a weak correlation between, say,
SAT scores and 'scholastic aptitude', and such a test might have utility when you
are trying to screen thousands of people for thousands of openings.
Others have pointed out that Lisp projects tend toward small numbers of
programmers as opposed to cadres of mercenaries that Java encourages.
I would argue that no arbitrary third-party certification test is going to give you
as high a correlation to the skills that you are looking for as a small panel of
competent people asking questions, face-to-face, that they think are relevant to
their needs.
Reliance on such a test simply shows that there is no one of suitable competence
in an organization to either judge the worth (or lack thereof) of such a test or
devise their own test that correlates better to their own needs.
It seems to me that reliance on mass statistical analysis in this case is losing
vis-a-vis common-sense interpersonal judgement and intuition.
But that's just my intuition.
-James Russell
* J L Russell
| Others have pointed out that Lisp projects tend toward small numbers of
| programmers as opposed to cadres of mercenaries that Java encourages.
This means that you need a lot more than your regular certification.
Which is why I think giving a paper at a conference is a good start.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
"Tim Bradshaw" <···@cley.com> wrote in message ····················@cley.com...
> * J L Russell wrote:
> > And all I'm saying is, do we really want to work for the forces of
> > darkness?
>
> We'd like to get paid, yes.
>
> --tim
So, Philip Morris, RJR, BAT, etc. approaches you (pl.) and says,
"Look, we know that lisp is the most effective tool around.
We want you to design a system that consolidates all our
research on increasing addictiveness of cigarettes and
marketing them to children in order to help us design
a more effective strategy."
You think, wow, we're so desperate, we'll take anything.
I don't know about you (pl.), but I'd rather go on the dole.
My utility to society would be _increased_ that way.
-James Russell
"J L Russell" <·········@alum.mit.edu> writes:
> I don't know about you (pl.), but I'd rather go on the dole.
> My utility to society would be _increased_ that way.
Would your landlord or mortgage lender reward you for this increase?
"Society" is not so appreciative of such individual sacrifice.
--
Steven E. Harris :: ········@raytheon.com
Raytheon :: http://www.raytheon.com
* J L Russell wrote:
> So, Philip Morris, RJR, BAT, etc. approaches you (pl.) and says,
> "Look, we know that lisp is the most effective tool around.
> We want you to design a system that consolidates all our
> research on increasing addictiveness of cigarettes and
> marketing them to children in order to help us design
> a more effective strategy."
> You think, wow, we're so desperate, we'll take anything.
So, IBM, Sun, Cisco etc. approaches you (pl.) and says,
"Look, we know that lisp is the most effective tool around.
We want you to design a system that consolidates all our
research on increasing reliability of systems and
marketing them to professionals in order to help us design
a more effective strategy."
You think, wow, how cool is that, we'll definitely take that.
Or, I guess, in your case `whoa, evil capitalist scum, stay back!'.
"Tim Bradshaw" <···@cley.com> wrote in message ····················@cley.com...
> * J L Russell wrote:
>
> > So, Philip Morris, RJR, BAT, etc. approaches you (pl.) and says,
> > "Look, we know that lisp is the most effective tool around.
> > We want you to design a system that consolidates all our
> > research on increasing addictiveness of cigarettes and
> > marketing them to children in order to help us design
> > a more effective strategy."
> > You think, wow, we're so desperate, we'll take anything.
>
> So, IBM, Sun, Cisco etc. approaches you (pl.) and says,
> "Look, we know that lisp is the most effective tool around.
> We want you to design a system that consolidates all our
> research on increasing reliability of systems and
> marketing them to professionals in order to help us design
> a more effective strategy."
> You think, wow, how cool is that, we'll definitely take that.
>
Point. Well, sure, it goes both ways. I'm not going to argue that.
On the other hand, I would want to ask what are the ultimate goals
of this company, at what cost, and how much would I be helping to
further them. Do they want to make a profit by selling useful
goods and services, or by poisoning people? Do they sell latter-day
Hollerith tabulators to latter-day Nazis? Do they aim to become a
monopoly and subvert the good aspects of capitalism? Do they
aim to buy the organs of education and government and hold humanity
in thrall in their furturistic dystopia?
Sure, they haven't done a good job so far, but do I want to help them?
-James Russell
"J L Russell" <·········@alum.mit.edu> wrote in message ····························@bgtnsc04-news.ops.worldnet.att.net...
>
> "Tim Bradshaw" <···@cley.com> wrote in message ····················@cley.com...
> > * J L Russell wrote:
> >
> > > So, Philip Morris, RJR, BAT, etc. approaches you (pl.) and says,
> > > "Look, we know that lisp is the most effective tool around.
> > > We want you to design a system that consolidates all our
> > > research on increasing addictiveness of cigarettes and
> > > marketing them to children in order to help us design
> > > a more effective strategy."
> > > You think, wow, we're so desperate, we'll take anything.
> >
> > So, IBM, Sun, Cisco etc. approaches you (pl.) and says,
> > "Look, we know that lisp is the most effective tool around.
> > We want you to design a system that consolidates all our
> > research on increasing reliability of systems and
> > marketing them to professionals in order to help us design
> > a more effective strategy."
> > You think, wow, how cool is that, we'll definitely take that.
> >
> Point. Well, sure, it goes both ways. I'm not going to argue that.
> On the other hand, I would want to ask what are the ultimate goals
> of this company, at what cost, and how much would I be helping to
> further them. Do they want to make a profit by selling useful
> goods and services, or by poisoning people? Do they sell latter-day
> Hollerith tabulators to latter-day Nazis? Do they aim to become a
> monopoly and subvert the good aspects of capitalism? Do they
> aim to buy the organs of education and government and hold humanity
> in thrall in their furturistic dystopia?
> Sure, they haven't done a good job so far, but do I want to help them?
>
> -James Russell
>
Listen, I don't want to be a hypocrite here, and I'm not trying to be or
accusing anyone else of it. I'm just using exaggerated examples for the
sake of argument, and I no longer have the time or will to argue.
My real life example:
I often, out of convenience, go to Burger King and get a veggie burger.
I'm not directly contributing to the death of any animals, but of course I'm
increasing the profits of a company that regularly traffics in dead animals.
I'm helping them expand their capital base and increase the traffic in dead
animals. I'm not in a position to grow my own vegetables on any meaningful
scale. I'm not willing to restructure my life to buy and eat food from
completely vegetarian companies.
We all make compromises, and I don't mean to impugn anyone.
Anyway, I guess I really don't have a point that is worth shedding any more
virtual ink over, so I'll shut up now.
My apologies for wasting everyone's time.
-James Russell
* J L Russell wrote:
> Point. Well, sure, it goes both ways. I'm not going to argue that.
> On the other hand, I would want to ask what are the ultimate goals
> of this company, at what cost, and how much would I be helping to
> further them. [...]
Well, yes we all want to do that. But that's kind of orthogonal to
the issue about Lisp I think. If Lisp becomes more widely used
(though, of course, it's more widely used than lots of people think
already), then big companies use it, and our choices are now bigger.
Right now (again, parodying a bit), the situation is that you can work
for a big company, or a small company, or a good company or a bad
company (those are meant to be two independent bits of information),
and you can get a job writing Java or C++. We want it to be the case
that you can work for good/bad & big/small, using x/y/CL/z/... This
is just an increase in the choice you have - you *already* need to
decide if you want to work for scum or the good guys, but I want it to
be the case that you can work for scum or the good guys *in Lisp*.
Ideally of course, I'd like it to be the case that you only get to
write Lisp if you work for the good guys, but I don't think that can
be arranged...
--tim
Tim Bradshaw wrote:
> * J L Russell wrote:
>
>>Point. Well, sure, it goes both ways. I'm not going to argue that.
>>On the other hand, I would want to ask what are the ultimate goals
>>of this company, at what cost, and how much would I be helping to
>>further them. [...]
>
>
> Well, yes we all want to do that. But that's kind of orthogonal to
> the issue about Lisp I think. If Lisp becomes more widely used
> (though, of course, it's more widely used than lots of people think
> already), then big companies use it, and our choices are now bigger.
Yeah. I am jumping in here just to remind everyone that this came up at
a conference during which the issue of Lisp's popularity came up again
and again. And again. Then there was a fishbowl on it. Then we broke for
snacks and talked about it.
So "do we want F500 acceptance" was a given, we do not need to bat that
around. This is not to say that some of us would rather starve than work
for f500 (or get certified or take a lie-detector test or give a urine
sample or whatever), just that enough Lispers want f500 acceptance that
we can now look at a second question: how much would certification
matter? We have one data point from Mr. megaCorp. Any others?
At the same time, i suddenly find myself intrigued by the question of
what would we want on a Lisp certfication exam?
(without-mush-thought...
(with-overlap...
-- destructively resplice this list to sort in place
-- (lambda ...
-- macrology
-- special variables
-- closures
-- CLOS
-- &key &rest &aux
-- not loop :)
-- lexical vs dynamic scope
-- some traps for those who would over-cons (append when nconc would do)
-- unwind-protect
-- exceptions))
OK, that looks like a TOC for any of the Lisp books out there, but I
guess the fun starts when we start writing the actual exam questions and
set out to make them really really nasty.
The neat things is that this might tie in nicely with someone's "to hell
with certification, let's work out how to teach Lisp".
aside: wasn't "The C Puzzle Book" by feuer great?
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
Kenny Tilton <·······@nyc.rr.com> writes:
> Tim Bradshaw wrote:
> > * J L Russell wrote:
> >
>
> >>Point. Well, sure, it goes both ways. I'm not going to argue that.
> >>On the other hand, I would want to ask what are the ultimate goals
> >>of this company, at what cost, and how much would I be helping to
> >>further them. [...]
> > Well, yes we all want to do that. But that's kind of orthogonal to
>
> > the issue about Lisp I think. If Lisp becomes more widely used
> > (though, of course, it's more widely used than lots of people think
> > already), then big companies use it, and our choices are now bigger.
>
> Yeah. I am jumping in here just to remind everyone that this came up
> at a conference during which the issue of Lisp's popularity came up
> again and again. And again. Then there was a fishbowl on it. Then we
> broke for snacks and talked about it.
This looks like a classic case; it seems Lisp has a self-esteem problem.
Why do we tend to beat ourselves up for what we think we are not? I
will take a chance on being called sexist, and say that many of the
women I know (including my wife and daughters) don't think highly of
their own appearence. Why is that? I believe it is because of the
standard they and the rest of society hold themselves up to, namely
the advertising media. What is more ironic is that many of the models
that are put up on the TV screen lead even more desparate lives,
not even living up to the standard their personae set when up on
the tube, with all of its touch-ups and tricks of photography.
How does this relate to Lisp? Lisp is a beautiful woman, secretly
adored by many F500 companies (but which some despise for whoring
with 1980's AI hype, and so which none of the F500 companies would
dare to admire openly). Does she get a bad rep from such past mistakes?
Absolutely. Would certification make any difference in her reputation?
I doubt it.
> So "do we want F500 acceptance" was a given, we do not need to bat
> that around. This is not to say that some of us would rather starve
> than work for f500 (or get certified or take a lie-detector test or
> give a urine sample or whatever), just that enough Lispers want f500
> acceptance that we can now look at a second question: how much would
> certification matter? We have one data point from Mr. megaCorp. Any
> others?
I think Lisp should stop looking to external sources to find her
self-acceptance, and instead look inward for what she knows is there.
She should find things that really do need improving, and improved
them, taking comfort even during that change in the knowledge that
she is beautiful in spite of her imperfections.
Self improvement is important. But without self-esteem, self-improvement
will never follow.
--
Duane Rettig ·····@franz.com Franz Inc. http://www.franz.com/
555 12th St., Suite 1450 http://www.555citycenter.com/
Oakland, Ca. 94607 Phone: (510) 452-2000; Fax: (510) 452-0182
In article <·············@beta.franz.com>, Duane Rettig <·····@franz.com> wrote:
> Lisp is a beautiful woman,
Interesting metaphor.
> secretly adored by many F500 companies
I'd say not-so-secretly adored by a few individuals who happen to be
working at F500 companies, but the companies themselves (whith is to say
the people who actually run the companies) treat her with attitudes
ranging from indifference to contempt. (Which attitudes are not wholly
unjustifiable IMO.)
> Would certification make any difference in her reputation?
> I doubt it.
I agree with that.
> She should find things that really do need improving, and improve them
In my experience, people who try to point out things that need improving
are not well received by the Lisp community (at least not as represented
by this newsgroup).
> she is beautiful in spite of her imperfections.
Yes, indeed.
> Self improvement is important. But without self-esteem, self-improvement
> will never follow.
Yes, I think that is really the nub of the matter.
E.
···@jpl.nasa.gov (Erann Gat) writes:
> In article <·············@beta.franz.com>, Duane Rettig <·····@franz.com> wrote:
>> Lisp is a beautiful woman,
>> secretly adored by many F500 companies
>
> I'd say not-so-secretly adored by a few individuals who happen to be
> working at F500 companies, but the companies themselves (whith is to say
> the people who actually run the companies) treat her with attitudes
> ranging from indifference to contempt. (Which attitudes are not wholly
> unjustifiable IMO.)
Well, I don't know how different it is in the US, but in the UK you
could replace "Lisp" in the above with "IT in general" and it would
still be about as true. "The people who actually run the companies"
are probably not the people whose perceptions of Lisp we (for whatever
value of "we") should be looking to change.
-dan
--
http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources
From: Kenny Tilton
Subject: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <3DC6F81A.2050609@nyc.rr.com>
Duane Rettig wrote:
> Kenny Tilton <·······@nyc.rr.com> writes:
>>Yeah. I am jumping in here just to remind everyone that this came up
>>at a conference during which the issue of Lisp's popularity came up
>>again and again. And again. Then there was a fishbowl on it. Then we
>>broke for snacks and talked about it.
>
>
> This looks like a classic case; it seems Lisp has a self-esteem problem.
Check. I wish you'd been there for the later "liquor bowl" where I got
that off my chest.
> Would certification make any difference in her reputation?
> I doubt it.
Well, I am guessing you would know as part of an organization that must
get a lot of feedback along these lines. otoh, if (if!) it is a
no-brainer to put cert process together, why not knock off one more
objection? but if it was just one (or a few) manager's insight,
fuggedaboutit.
> I think Lisp should stop looking to external sources to find her
> self-acceptance, and instead look inward for what she knows is there.
> She should find things that really do need improving, and improved
> them, ...
Right. At ALU 2003 it should be cream pies in the face for anyone
whining about Lisp not being popular, or suggesting we all switch to
JavaScript since it's pretty close to Lisp.
I told my Franz rep y'all should send a thick-skinned, genial,
proficient Lisper who is also Python-proficient into c.l.python to
harrass them regularly with "why don't you use Lisp?". Anytime they lose
the argument, Franz knows what to build next.
Popularity is great since it means more Lisp jobs, more Lisp open source
projects, more revenue for Lisp vendors to reinvest. But Lisp won't get
more popular running around with its tail between its legs, such as by
running away from the syntax a la Dylan. That just makes us look pathetic.
Ever see an objectively homely person surrounded by adoring admirers?
The key is confidence, geniality, wit, good humor, etc etc.
Build it and they will come.(tm) Hell, considering Python and Ruby and
even Java, they are already on the way, they just don't know it. When
they get to our doors we should turn them away three times before
letting them in.
And when someone asks what language we use, the answer should be, "Lisp.
What else?" or "Lisp. Don't you?"
"I am Lisper, hear me roar,
With productivity too great to ignore..."
Prior art: Erik tried to whip up (paraphrasing) a "say it loud, say it
proud" mentality many moons ago. What he said.
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
From: Fred Gilham
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <u7lm4827cr.fsf@snapdragon.csl.sri.com>
Well, now that Franz in the person of Duane Rettig has jumped into the
fray, I should say that I talked to the telephone guy at the
conference, not just once, but 3 or 4 times. My feeling was that he
was exactly the kind of customer Franz wants. Why? He didn't just
want to go to some software shop and get a shrink-wrapped package. He
wanted training, hand-holding and support.
The problem was that when I pointed out to him that he should just
*talk* to someone from Franz, he was very resistant. He knew what he
wanted, and nothing we had was it. Eventually I decided that he
lacked imagination. He kept repeating that he didn't know how to
program; my thought was that he should have brought a programmer with
him --- either the guy who was interested in Lisp, or, perhaps better,
someone else who didn't know anything about it but who knew
programming....I know that's not REALLY possible but we must make
allowances.... :-)
Anyway I wasn't quite ready to label him as pointy-haired, but close
to it. So perhaps we can't judge everything by him.
--
Fred Gilham ······@csl.sri.com
Jordan Hubbard: We have a crash bug. It needs to be fixed. We DO NOT
need to know how to print 3000 spaces in 11 different languages! :-)
Daniel Sobral: I concur. But if anyone wants to do it with loader,
: 3kbl 3000 0 do bl emit loop ; 3kbl will do the trick.
From: Greg Menke
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <m3vg3cokc7.fsf@europa.pienet>
Fred Gilham <······@snapdragon.csl.sri.com> writes:
>
> The problem was that when I pointed out to him that he should just
> *talk* to someone from Franz, he was very resistant. He knew what he
> wanted, and nothing we had was it. Eventually I decided that he
> lacked imagination. He kept repeating that he didn't know how to
> program; my thought was that he should have brought a programmer with
> him --- either the guy who was interested in Lisp, or, perhaps better,
> someone else who didn't know anything about it but who knew
> programming....I know that's not REALLY possible but we must make
> allowances.... :-)
I spoke with him too. He seemed to have some particularly fixed ideas
about what the Lisp Community should do as far as making Lisp more
marketable to "Fortune 500" people. I think he wants Lisp community
action as opposed to talking to a vendor- but I'm not sure why he
wasn't interested hooking up with a vendor.
I do give him credit for going to the trouble of showing up though.
> Anyway I wasn't quite ready to label him as pointy-haired, but close
> to it. So perhaps we can't judge everything by him.
He does believe in the management fads; 6 Sigma being one he mentioned
as "up and coming".
Gregm
From: Michael Hudson
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <7h3znso179r.fsf@pc150.maths.bris.ac.uk>
Kenny Tilton <·······@nyc.rr.com> writes:
> I told my Franz rep y'all should send a thick-skinned, genial,
> proficient Lisper who is also Python-proficient into c.l.python to
> harrass them regularly with "why don't you use Lisp?".
All that would do is fuck people off, I'd be willing to bet. People
don't like to be told they're doing things wrong, even if they are.
Cheers,
M.
--
> It might get my attention if you'd spin around in your chair,
> spoke in tongues, and puked jets of green goblin goo.
I can arrange for this. ;-) -- Barry Warsaw & Fred Drake
From: Paul F. Dietz
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <xG6dnUPyaOaZUVqgXTWc3w@dls.net>
Michael Hudson wrote:
> All that would do is fuck people off, I'd be willing to bet. People
> don't like to be told they're doing things wrong, even if they are.
They might appreciate a Python-the-language implementation on Lisp,
though. It would be interesting to compare performance.
Paul
From: Michael Hudson
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <7h3n0on27nr.fsf@pc150.maths.bris.ac.uk>
"Paul F. Dietz" <·····@dls.net> writes:
> Michael Hudson wrote:
>
> > All that would do is fuck people off, I'd be willing to bet. People
> > don't like to be told they're doing things wrong, even if they are.
>
> They might appreciate a Python-the-language implementation on Lisp,
> though. It would be interesting to compare performance.
Oh yes, I'm not saying that any mention of lisp is inflammatory, it's
just that swanning in and demanding to know why everyone's not using
lisp would be counter-productive (unless you want to produce hot air).
Answering questions and putting in asides like "hmm, this would be
easier in CL" seems a better way of engaging people's curiosity.
And Python-in-lisp would indeed be interesting (Lython, anyone?).
Cheers,
M.
--
ARTHUR: Yes. It was on display in the bottom of a locked filing
cabinet stuck in a disused lavatory with a sign on the door
saying "Beware of the Leopard".
-- The Hitch-Hikers Guide to the Galaxy, Episode 1
From: Ng Pheng Siong
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <aqe063$eqo$1@reader01.singnet.com.sg>
According to Michael Hudson <···@python.net>:
> And Python-in-lisp would indeed be interesting (Lython, anyone?).
The better name is "Hikk", like, a python hissing with a lisp.
(Well, I don't know if pythons hiss or not. ;-)
--
Ng Pheng Siong <····@netmemetic.com> * http://www.netmemetic.com
From: Tim Lavoie
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <87of91s2uq.fsf@theasylum.dyndns.org>
>>>>> "Ng" == Ng Pheng Siong <····@netmemetic.com> writes:
Ng> According to Michael Hudson <···@python.net>:
>> And Python-in-lisp would indeed be interesting (Lython,
>> anyone?).
Ng> The better name is "Hikk", like, a python hissing with a lisp.
Ng> (Well, I don't know if pythons hiss or not. ;-)
I dunno, "Lython" thounds lithpy to me.
--
"Why does man kill? He kills for food. And not only food: frequently
there must be a beverage."
--Woody Allen
Michael Hudson <···@python.net> writes:
> Oh yes, I'm not saying that any mention of lisp is inflammatory, it's
> just that swanning in and demanding to know why everyone's not using
> lisp would be counter-productive (unless you want to produce hot air).
Obviously, since if any lisp vendor actually used python or perl or C,
they'd know what the community of Not-A-Lisp-Machine-and-Happy-About-It
users wants and needs and wouldn't have to waste our time asking.
Siskind's Stalin compiler is The Right Thing. Too bad it's such a
notable rarity.
Scott Schwartz <··········@usenet ·@bio.cse.psu.edu> writes:
> Siskind's Stalin compiler is The Right Thing. Too bad it's such a
> notable rarity.
The object code produced by a delivery compiler such as Stalin is
unsuitable for patching in the field and often fatally differs from
your source code. Trading debugging flexibility for small increases
in runtime performance is a splendid way to lose.
There have been numerous delivery-only compilers for Lisp which have
quietly come and gone. I suspect most users of Lisp accept that the
software debugging cycle is never truly over; you are always debugging
your application, even after delivery.
Scott Schwartz <··········@usenet ·@bio.cse.psu.edu> writes:
> Michael Hudson <···@python.net> writes:
> > Oh yes, I'm not saying that any mention of lisp is inflammatory, it's
> > just that swanning in and demanding to know why everyone's not using
> > lisp would be counter-productive (unless you want to produce hot air).
>
> Obviously, since if any lisp vendor actually used python or perl or C,
> they'd know what the community of Not-A-Lisp-Machine-and-Happy-About-It
> users wants and needs and wouldn't have to waste our time asking.
>
> Siskind's Stalin compiler is The Right Thing. Too bad it's such a
> notable rarity.
Say what? Python's (and for that matter Common Lisp's) interactive
development is one of the thing it has going for it...
Cheers,
M.
--
This proposal, if accepted, will probably mean a heck of a lot of
work for somebody. But since I don't want it accepted, I don't
care. -- Laura Creighton, PEP 666
From: Tim Bradshaw
Subject: Re: Lisp Machines considered Inferior
Date:
Message-ID: <ey3u1ipqy77.fsf@cley.com>
* Scott Schwartz wrote:
> Siskind's Stalin compiler is The Right Thing. Too bad it's such a
> notable rarity.
whole-program analysis is a nice way of ensuring your compile times
compete poorly even with typical C++ compilers. The right thing, if
your time is free and you don't like things like runtime patching,
maybe.
--tim
Tim Bradshaw <···@cley.com> writes:
> whole-program analysis is a nice way of ensuring your compile times
> compete poorly even with typical C++ compilers.
If you don't like Stalin, have an option to do less optimization. But
it's not optional to produce small standalone native a.out files.
I don't want a 128MB runtime system in my 128 line program. I
definately don't want a 128MB runtime system in my application that
uses 999MB of RAM. I don't want a Lisp Machine emulator. I don't
want lisp to be my shell. I want a compiler that works like gcc, and
applications that work like normal unix applications.
That's the only worthwile use of my time. Anything else is just proof
that lisp is unusable by serious unix programmers.
It's really sad. Half the message here are folks trying to figure out
why lisp isn't popular. The other half is them refusing to belive it
when told.
I just filed a bug report for CLISP (for the second time). Nothing
complicated: if a non-interactive program gets a certain kind of
error, it pops into an interactive break loop, writing garbage to
standard output and consuming standard input and not quitting when you
hit ctrl-C. That's anti-unix engineering to an unthinkable degree.
It would be inconceiveable for perl's runtime system to produce
garbage like that. And this is supposed to be one of the best lisp
implementations? I belive it.
Scott Schwartz <··········@usenet ·@bio.cse.psu.edu> writes:
> Tim Bradshaw <···@cley.com> writes:
> > whole-program analysis is a nice way of ensuring your compile times
> > compete poorly even with typical C++ compilers.
>
> If you don't like Stalin, have an option to do less optimization. But
> it's not optional to produce small standalone native a.out files.
On the assumption that you're not just trolling... why? And just how
standalone is a "native a.out" file?
> I don't want a 128MB runtime system in my 128 line program. I
> definately don't want a 128MB runtime system in my application that
> uses 999MB of RAM. I don't want a Lisp Machine emulator. I don't
> want lisp to be my shell. I want a compiler that works like gcc, and
> applications that work like normal unix applications.
How robust is a normal unix application? How well does it handle
unexpected input? What about bit errors? How easy is in-place
maintenance? How long a downtime does making major changes to the
system entail? And if your application uses 999MB of RAM, I'd guess
that the cost of an extra 512MB stick is lost in the noise, even
assuming that this runtime isn't being shared by any other process on
the system.
> That's the only worthwile use of my time. Anything else is just proof
> that lisp is unusable by serious unix programmers.
There's a rather obvious flaw in this syllogism that I will refrain
from commenting on further.
> [ CLISP and anti-unix ]
> And this is supposed to be one of the best lisp
> implementations? I belive it.
That's nice.
Your attitude thus far seems to be that of someone who has only ever
seen hammers; then, given a Swiss Army knife, expresses frustration
when banging nails in seems to be harder. There are other ways of
assembling useful objects...
Christophe
Frivolous Unix programmer
--
http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)
Christophe Rhodes <·····@cam.ac.uk> writes:
> > If you don't like Stalin, have an option to do less optimization. But
> > it's not optional to produce small standalone native a.out files.
>
> On the assumption that you're not just trolling... why?
Why would you want anything else? Smaller is better, because you
don't waste space. On a real system you will have hundreds or
thousands of independent processes in execution, so it would be bad to
have them burn more memory than necessary. On a real system you want
to take advantage of the system's basic mechanisms, so that means
producing native format executables.
For example, suppose you want to profile your program. You've got a
MIPS or Alpha so you say something like, "pixie a.out" and then run
"a.out.pixie" and you get your results. Can't do that with some
random byte code, or non-system binary.
> And just how standalone is a "native a.out" file?
I don't like shared libraries, but I guess they're a fact of life
these days.
> How robust is a normal unix application?
Look at, say, qmail. Totally bullet proof.
> How easy is in-place maintenance?
Very. Just load a new rpm when the time comes.
> How long a downtime does making major changes to the
> system entail?
Mere moments. You won't even have time to fire up emacs. :)
> And if your application uses 999MB of RAM, I'd guess
> that the cost of an extra 512MB stick is lost in the noise,
Sure, if we had free slots, but when the current system (1024
processors with 1GB each) was purchased, we couldn't fit more than
that. If we could, we'd run bigger jobs, and still not have any RAM
to waste.
> even assuming that this runtime isn't being shared by any other process on
> the system.
No, that would still be no excuse.
> Your attitude thus far seems to be that of someone who has only ever
> seen hammers; then, given a Swiss Army knife, expresses frustration
> when banging nails in seems to be harder. There are other ways of
> assembling useful objects...
On the contrary. I've used and enjoyed a variety of systems. But I
know a spork when I see it. (And perl is the canonical swiss army
knife.)
* Scott Schwartz
| Smaller is better, because you don't waste space. On a real system you
| will have hundreds or thousands of independent processes in execution, so
| it would be bad to have them burn more memory than necessary.
This must be why the Linux kernels I build warn me that they too big for
the floppy disk, even when they are compressed.
| Look at, say, qmail. Totally bullet proof.
*laugh*
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> writes:
> * Scott Schwartz
> | Smaller is better, because you don't waste space. On a real system you
> | will have hundreds or thousands of independent processes in execution, so
> | it would be bad to have them burn more memory than necessary.
>
> This must be why the Linux kernels I build warn me that they too big for
> the floppy disk, even when they are compressed.
Sure, 1.44 MB is a lot. Linux at least has the excuse of supporting
lots of devices. On a very vanilla system a kernel of a few hundred
KB is easily managed.
How big did you say your lisp runtime was?
> | Look at, say, qmail. Totally bullet proof.
>
> *laugh*
If you can reimplement qmail in lisp, and do as well, I'd be amazed
and delighted.
* Scott Schwartz
| Sure, 1.44 MB is a lot. Linux at least has the excuse of supporting lots
| of devices. On a very vanilla system a kernel of a few hundred KB is
| easily managed.
When built, /usr/src/linux is 3,569,040 bytes and /boot/vmlinuz is
1,303,242 bytes. I have no idea why it still complains about the floppy
size, it should have fit. However, my machine now has 1,048,576K of
memory, but systemt status reports say only 1,032,908K are available. As
far as I can see, that is 15,668K of memory in the kernel alone.
| How big did you say your lisp runtime was?
The full Common Lisp environment with a lot of stuff pre-loaded consumes
all of 8,576K of memory when it starts up. That is half the Linux kernel.
| If you can reimplement qmail in lisp, and do as well, I'd be amazed and
| delighted.
WTF would I reimplement such a braindamaged design for?
Dude, your whole weltanschauung needs a serious readjustment. You have
walked too far in the one direction you have chosen to see that any other
direction you could have chosen would have brought you to places with no
loss of convenience or quality of life, even if you had not walked as far,
but from where you are now to any other such place is probably farther
than it would have been from the common starting point. This is how
people get trapped in the Valley of Death with Microsoft, too, for the
farther they walk into that valley, the taller the mountains they would
have to scale to get out of it. I always saw Unix as the Great Plains.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
* Erik Naggum <····@naggum.no>:
> However, my machine now has 1,048,576K of memory, but systemt status
> reports say only 1,032,908K are available. As far as I can see,
> that is 15,668K of memory in the kernel alone.
That's nothing. My kernel uses nearly half my memory for this thing
top calls "cache". It's an outrage.
--
Johannes Gr�dem <OpenPGP: 5055654C>
* "Johannes Gr�dem" <······@ifi.uio.no>
| That's nothing. My kernel uses nearly half my memory for this thing
| top calls "cache". It's an outrage.
Yeah, tell me about it. I upgraded from 524,288K to 1,048,576K because
all the memory was in use, but after I upgraded, all the memory is still
in use. And I am quite sure this is because of The Great Common Lisp
Conspiracy to Consume All Available Memory. I think Common Lisp in space
failed because there are far less atoms in space than here on earth, so
when it started to consume atoms to use for memory, the Universe was in
danger of being eaten up by it and the project just had to be aborted.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum wrote:
> * "Johannes Gr�dem" <······@ifi.uio.no>
> | That's nothing.��My�kernel�uses�nearly�half�my�memory�for�this�thing
> | top calls "cache".��It's�an�outrage.
>
> Yeah,�tell�me�about�it.��I�upgraded�from�524,288K�to�1,048,576K�because
> all�the�memory�was�in�use,�but�after�I�upgraded,�all�the�memory�is�still
> in�use.
This is a common newbie question[1]. I'm shocked to learn that none of the
40+ O'Reilly books you own clued you in on this! Why don't you practice
what you preach and RTFM?
F/up2 linux.misc
Oleg
[1] Linux tries to use as much RAM as possible. That's what it's there for.
http://www.tldp.org/FAQ/Linux-FAQ/common-problems.html#FREE-MEMORY-KEEPS-SHRINKING
* Johannes Gr�dem
| That's nothing.��My�kernel�uses�nearly�half�my�memory�for�this�thing
| top calls "cache".��It's�an�outrage.
* Erik Naggum
| Yeah,�tell�me�about�it.��I�upgraded�from�524,288K�to�1,048,576K�because
| all�the�memory�was�in�use,�but�after�I�upgraded,�all�the�memory�is�still
| in�use.
* Oleg <············@myrealbox.com>
| This is a common newbie question[1]. I'm shocked to learn that none of
| the 40+ O'Reilly books you own clued you in on this! Why don't you
| practice what you preach and RTFM?
Have you considered the option of getting the joke? If not, try it now
and redeem your soul.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> writes:
> than it would have been from the common starting point. This is how
> people get trapped in the Valley of Death with Microsoft, too, for the
> farther they walk into that valley, the taller the mountains they would
> have to scale to get out of it.
No, it's because their brains are washed by the local priests until
they actually believe that the world ends at thouse mountain ridges.
--
(espen)
I would like to congratulate you on bringing a smile to my face on an
otherwise rather grey and drab winters evening.
Good luck in your quest.
:)w
From: Tim Bradshaw
Subject: Re: Lisp Machines considered Inferior
Date:
Message-ID: <ey3heeprnov.fsf@cley.com>
I wish I could followup to two threads at once.
* Scott Schwartz wrote:
> On a real system you want to take advantage of the system's basic
> mechanisms, so that means producing native format executables.
* Andre van Meulebrouck wrote:
> For instance, .Net isn't language centric; how about targeting .Net
> with CL?
Now one of these has to be wrong...
(and incidentally:
[···@penzance tfb]$ file /local/bin/clc
/local/bin/clc: ELF 32-bit LSB executable, Intel 80386, version 1, dynamically linked (uses shared libs), not stripped
[···@penzance tfb]$ ls -l /local/bin/clc
-rwxrwxrwx 1 tfb tfb 3342364 Jul 21 11:28 /local/bin/clc
That's a native format executable. Made with Lisp. Usable from the
Unix command line, starup time of a few hundredths of a second. It's
3MB because it doesn't use shared libraries other than libc (which
Scott Schwartz abhors, so he'll be pleased it doesn't) and it uses
CORBA.)
--tim
Tim Bradshaw <···@cley.com> writes:
> /local/bin/clc: ELF 32-bit LSB executable, Intel 80386, version 1, dynamically linked (uses shared libs), not stripped
> [···@penzance tfb]$ ls -l /local/bin/clc
> -rwxrwxrwx 1 tfb tfb 3342364 Jul 21 11:28 /local/bin/clc
>
> That's a native format executable.
It's huge. And world writable. What's it do?
From: Tim Bradshaw
Subject: Re: Lisp Machines considered Inferior
Date:
Message-ID: <ey3of8xrsma.fsf@cley.com>
* Scott Schwartz wrote:
> I don't want a 128MB runtime system in my 128 line program. I
> definately don't want a 128MB runtime system in my application that
> uses 999MB of RAM.
I think you are confusing Lisp with that notoriously unpopular
language, Java. Or maybe with some figment of your imagination, who
can tell?
--tim
* Scott Schwartz
| I don't want a 128MB runtime system in my 128 line program.
You are about two orders of magnitude off, and this only makes you look
incredibly ignorant and arrogant.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
> * In message <··············@galapagos.cse.psu.edu>
> * On the subject of "Re: Lisp Machines considered Inferior"
> * Sent on 10 Nov 2002 15:50:35 -0500
> * Honorable Scott Schwartz <··········@usenet ·@bio.cse.psu.edu> writes:
>
> it's not optional to produce small standalone native a.out files.
this has been rehashed many times here - suffice it to say that
the word "standalone" is largely meaningless (ever since libc.so
appeared on the scene or even before that).
> I just filed a bug report for CLISP (for the second time). Nothing
> complicated: if a non-interactive program gets a certain kind of
> error, it pops into an interactive break loop, writing garbage to
> standard output and consuming standard input and not quitting when you
> hit ctrl-C. That's anti-unix engineering to an unthinkable degree.
how many programs would behave well when given /dev/zero as STDIN?
(that's what you used as the standard input - /dev/zero, _not_ /dev/null)
The small bug you did uncover has been already fixed.
--
Sam Steingold (http://www.podval.org/~sds) running RedHat8 GNU/Linux
<http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/>
<http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html>
I don't want to be young again, I just don't want to get any older.
Sam Steingold <···@gnu.org> writes:
> how many programs would behave well when given /dev/zero as STDIN?
All correct ones. The one I sent you didn't even read from stdin,
and there was no reason for CLISP to do so.
In response to a deliberately induced stack overflow in a
non-interactive program (in other words, a minimal example constructed
to demonstrate a large problem), clisp jumped into an interactive
break loop and started talking to stdin/stdout. The only correct
behaviour would have been to print a message to stderr and to exit,
because there was no terminal, no user, no possibility for
interaction. If stdin had been connected to a network server, this
would have been a security problem of the highest order. The fact
that the break loop was boggled by ascii NUL characters is wierd, but
secondary, given that it had no right to be looking at them in the
first place.
> (that's what you used as the standard input - /dev/zero, _not_ /dev/null)
Of course. It was a convenient source of input, to drive home the
point that CLISP is reading stdin, when it stdin is connected to
arbitrary user-supplied input.
> The small bug you did uncover has been already fixed.
Not by the patch you sent me. CLISP still writes garbage to standard
output instead of standard error. It still assumes that debug_io is
open and usable. It still assumes that the runtime system can read
from standard-input whenever it wants to.
$ cat /etc/motd | clisp silent.cl 2>/dev/null | sed 's/.*/ERROR/'
ERROR
ERROR
From: Tim Bradshaw
Subject: Re: Lisp Machines considered Inferior
Date:
Message-ID: <ey3n0obayb6.fsf@cley.com>
* Scott Schwartz wrote:
> In response to a deliberately induced stack overflow in a
> non-interactive program (in other words, a minimal example constructed
> to demonstrate a large problem), clisp jumped into an interactive
> break loop and started talking to stdin/stdout. The only correct
> behaviour would have been to print a message to stderr and to exit,
> because there was no terminal, no user, no possibility for
> interaction. If stdin had been connected to a network server, this
> would have been a security problem of the highest order. The fact
> that the break loop was boggled by ascii NUL characters is wierd, but
> secondary, given that it had no right to be looking at them in the
> first place.
Wouldn't it have been slightly better if your program had a catch-all
error handler wrapped around it? It seems strange to complain about
weird behaviour from an implementation when you haven't bothered to
handle the error. Something like:
(handler-bind
((error (lambda (e)
(format *error-output* "~&Uncaught error: ~A~%")
(backtrace *error-output*)
#-com.cley/development
(exit-lisp 1))))
(top-level))
is what I use.
--tim
Hello!
Tim Bradshaw <···@cley.com> wrote:
>* Scott Schwartz wrote:
>> Siskind's Stalin compiler is The Right Thing. Too bad it's such a
>> notable rarity.
>whole-program analysis is a nice way of ensuring your compile times
>compete poorly even with typical C++ compilers. The right thing, if
>your time is free and you don't like things like runtime patching,
>maybe.
In fact, the whole-program optimizing compiler SmallEiffel is rather
fast, compilation times are dominated by the gcc run times (SmallEiffel
targets C). Even together, compiling an Eiffel program with SmallEiffel
plus gcc (in the C mode) is faster than compiling some C++ program
which heavily uses templates and such with gcc (in C++ mode).
Kind regards,
Hannah.
In article <··················@iapetus.uk.clara.net>, "Justin Johnson"
<·······@mobiusent.com> wrote:
>It almost sounds like if you created a C++ -> C converter for gcc, it would
>run faster than the native C++ compilation.
>
>If I remember correctly, the first C++ compilers used to work like that.
Hence their name: cfront
* Scott Schwartz
| Siskind's Stalin compiler is The Right Thing. Too bad it's such a
| notable rarity.
This is one of the major differences between the Common Lisp world and
the Scheme world. I am surprised how many people from the Scheme world
think that Common Lisp is no different from Scheme. It reminds me of C
people who believe that C++ is just a bloated C.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> writes:
> This is one of the major differences between the Common Lisp world and
> the Scheme world.
No, it's not. Like I said, Stalin is notably different.
> I am surprised how many people from the Scheme world
I'm not from the Scheme world.
* Scott Schwartz
| I'm not from the Scheme world.
Then you should check it out.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Kenny Tilton
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <3DC7EADB.E7726B1F@nyc.rr.com>
Michael Hudson wrote:
> Kenny Tilton <·······@nyc.rr.com> writes:
>
> > I told my Franz rep y'all should send a thick-skinned, genial,
> > proficient Lisper who is also Python-proficient into c.l.python to
> > harrass them regularly with "why don't you use Lisp?".
>
> All that would do is fuck people off, I'd be willing to bet. People
> don't like to be told they're doing things wrong, even if they are.
OK, I should not have said "harrass", that was meant ironically, and I do
not know of an emoticon for that.
"Why don't you use Lisp?" should only be asked when it really solves the
person's problem.
Folks may still take exception to well-meaning advice, but that would be
their problem, not the advice giver's. Any debate would either prove the
Lisper wrong (telling Franz what to do next) or drag on long enough to
persuade one Pythonian in a hundred to give Lisp a look.
Because you are right, most folks aren't looking to change languages at
the drop of a remark. btw, from what I hear the python NGs are pretty
cool, I am guessing that well-informed recommendations of Lisp would meet
with a considered and considerate reception.
kenny
clinisys
From: Tim Lavoie
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <87u1ivrh63.fsf@theasylum.dyndns.org>
>>>>> "KT" == Kenny Tilton <·······@nyc.rr.com> writes:
KT> Because you are right, most folks aren't looking to change
KT> languages at the drop of a remark. btw, from what I hear the
KT> python NGs are pretty cool, I am guessing that well-informed
KT> recommendations of Lisp would meet with a considered and
KT> considerate reception.
That sounds pretty close, I think, especially from some of the folks
who have been around long enough to appreciate where some of the
concepts in Python come from. They may not be inclined to switch, but
will pay attention.
I think that what would attract more useful attention in general from
a wide audience is a reasonably general-purpose, OS-portable "killer
app". Python has Zope; it runs under Unix systems and Windows, plays
nice with web servers, but can stand on its own too. Plenty of people
will try the app, then dig further under the hood and start to
appreciate what they find there.
This has helped to some degree where I work, in that Zope is accepted
for at least some projects, and Python development isn't high-profile,
but accepted as part of the package. I'm glad I get to do some, but
management still favours duct-tape J2EE work on top of bletcherous
third-party "solutions" for big contracts. They'd never dream of
paying to start something from scratch in CL, no matter how much the
geek with a CS degree raves about it.
Lisp needs that app which lots of people will use, at a level where
they can find it too. I see examples mentioned of telco switches, ATM
systems, air traffic control, and forensic analysis. All great stuff,
but most people aren't going to have access to those sorts of things,
never mind digging deeper into the application. We all know that Lisp
is great for hard problems, but it needs the attention of a broader
audience as well. *Sigh*
Well thanks all, I needed a rant today. :)
Tim
From: Chris Beggy
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <873cqf7o5m.fsf@lackawana.kippona.com>
Tim Lavoie <········@spamcop.net> writes:
> Lisp needs that app which lots of people will use, at a level where
> they can find it too. I see examples mentioned of telco switches, ATM
> systems, air traffic control, and forensic analysis. All great stuff,
> but most people aren't going to have access to those sorts of things,
> never mind digging deeper into the application. We all know that Lisp
> is great for hard problems, but it needs the attention of a broader
> audience as well. *Sigh*
Paul Graham has wisely chosen email *SPAM* as the first target
for his lispy ARC language. I presume it is for precisely the
reasons you have mentioned above.
Chris
From: Tim Lavoie
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <87pttjr9fh.fsf@theasylum.dyndns.org>
>>>>> "Chris" == Chris Beggy <······@kippona.com> writes:
Chris> Paul Graham has wisely chosen email *SPAM* as the first
Chris> target for his lispy ARC language. I presume it is for
Chris> precisely the reasons you have mentioned above.
Fair enough, his proposed solution seems to be getting some attention,
but it highlights the process, not the language. What I had in mind is
some package that people would mostly just install and run, but which
has user-visible Lisp in it. It should be small enough to become
common, but big enough to be non-trivial. Smaller than a locomotive,
but bigger than a breadbox.
From: Duane Rettig
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <43cqfpuhz.fsf@beta.franz.com>
Chris Beggy <······@kippona.com> writes:
> Tim Lavoie <········@spamcop.net> writes:
>
> > Lisp needs that app which lots of people will use, at a level where
> > they can find it too. I see examples mentioned of telco switches, ATM
> > systems, air traffic control, and forensic analysis. All great stuff,
> > but most people aren't going to have access to those sorts of things,
> > never mind digging deeper into the application. We all know that Lisp
> > is great for hard problems, but it needs the attention of a broader
> > audience as well. *Sigh*
>
> Paul Graham has wisely chosen email *SPAM* as the first target
> for his lispy ARC language. I presume it is for precisely the
> reasons you have mentioned above.
The "killer app" problem has been discussed before. It is not the
way to gain acceptance for a language, unless it _can't_ be done in
any other language (how many apps does that cover?) Graham's spam
examples are translated into lisp, he says, for greater accessibility.
The killer app has thus become greater than its langauge.
In general, once a killer app gains acceptance, its language is
forgotten or played down, and the app itself becomes more mundane.
Or, copycat apps grow up around it which may not have the pizazz
of the original, but are utilitarian nevertheless and remove the
"killer" label from the original. Examples of each: emacs,
macsyma -> mathematica.
--
Duane Rettig ·····@franz.com Franz Inc. http://www.franz.com/
555 12th St., Suite 1450 http://www.555citycenter.com/
Oakland, Ca. 94607 Phone: (510) 452-2000; Fax: (510) 452-0182
From: Tim Lavoie
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <87lm47qnrv.fsf@theasylum.dyndns.org>
>>>>> "Duane" == Duane Rettig <·····@franz.com> writes:
Duane> The "killer app" problem has been discussed before. It is
Duane> not the way to gain acceptance for a language, unless it
Duane> _can't_ be done in any other language (how many apps does
Duane> that cover?) Graham's spam examples are translated into
Duane> lisp, he says, for greater accessibility.
Duane> The killer app has thus become greater than its langauge.
Duane> In general, once a killer app gains acceptance, its
Duane> language is forgotten or played down, and the app itself
Duane> becomes more mundane. Or, copycat apps grow up around it
Duane> which may not have the pizazz of the original, but are
Duane> utilitarian nevertheless and remove the "killer" label from
Duane> the original. Examples of each: emacs, macsyma ->
Duane> mathematica.
Perhaps I overstated what I had hoped for; not acceptance in the
larger scale of things, so much as a way to get it far enough in the
door to stick around a while, and get some once the app is somewhat
familiar. Anyway, I apologise for beating what is apparently a dead
horse, I thought it might have twitched.
Duane Rettig:
* Or, copycat apps grow up around it which may not have the pizazz
* of the original, but are utilitarian nevertheless and remove the
* "killer" label from the original. Examples of each: emacs,
* macsyma -> mathematica.
this is quite unfair. wolfram had earlier (1983) written his own symbolic
mathematics package called SMP. i don't know if he saw macsyma at caltech
or not; we had both packages, and if memory serves, SMP had not a trace
of lisp or any part of macsyma in it. mathematica is based on his *own*
work, not a copycat of someone else's.
oz
---
music is the space between the notes. -- claude debussy
ozan s yigit <··@blue.cs.yorku.ca> writes:
> Duane Rettig:
>
> * Or, copycat apps grow up around it which may not have the pizazz
> * of the original, but are utilitarian nevertheless and remove the
> * "killer" label from the original. Examples of each: emacs,
> * macsyma -> mathematica.
>
> this is quite unfair. wolfram had earlier (1983) written his own symbolic
> mathematics package called SMP. i don't know if he saw macsyma at caltech
> or not; we had both packages, and if memory serves, SMP had not a trace
> of lisp or any part of macsyma in it. mathematica is based on his *own*
> work, not a copycat of someone else's.
I cannot say anything about SMP and Wolfram's early work, but, at
least until a few years ago, the guts of Mathematica (as exposed by
the kernel C interfacing subsystem) had a definite Lispy feeling.
Maybe things have changed now (I haven't touched the thing in years),
but that is that way I remember it.
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
715 Broadway 10th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
From: Jesper Harder
Subject: Re: mathematica [Re: I'm a Lisper, hear me roar... [etc]]
Date:
Message-ID: <m3znsmd6x6.fsf@defun.localdomain>
ozan s yigit <··@blue.cs.yorku.ca> writes:
> Duane Rettig:
>
> * macsyma -> mathematica.
>
> this is quite unfair. wolfram had earlier (1983) written his own symbolic
> mathematics package called SMP. i don't know if he saw macsyma at caltech
> or not;
He did know Macsyma. Maxima (DOE Macsyma) still includes a Macsyma
program written by Stephen Wolfram:
,----[ gamalg.usg ]
|
| Gamma Matrix Algebra Program (GAMALG).
| ======================================
|
| written by
| Anthony E Terrano (TREX)
| and Stephen Wolfram (SWOLF)
|
| mostly on [Feb 25, Mar 3], 1979
| revised July 10, 1979.
| [...]
| Further information, especially on the algorithms used by GAMALG, may
| be found in 'MACSYMA Tools for Feynman Diagram Calculations', by Stephen
| Wolfram, in Proceeding of the 1979 Users' Conf. and Caltech preprint
| CALT-68-720 (June 1979).
`----
From: Duane Rettig
Subject: Re: mathematica [Re: I'm a Lisper, hear me roar... [etc]]
Date:
Message-ID: <4iszamsu3.fsf@beta.franz.com>
ozan s yigit <··@blue.cs.yorku.ca> writes:
> Duane Rettig:
>
> * Or, copycat apps grow up around it which may not have the pizazz
> * of the original, but are utilitarian nevertheless and remove the
> * "killer" label from the original. Examples of each: emacs,
> * macsyma -> mathematica.
>
> this is quite unfair. wolfram had earlier (1983) written his own symbolic
> mathematics package called SMP.
Actually, according to web searches, he started it even earlier, in 1979.
But the Macsyma effort had started in 1967, long before he even started
using computers in 1973 (at age 14).
> i don't know if he saw macsyma at caltech
> or not; we had both packages, and if memory serves, SMP had not a trace
> of lisp or any part of macsyma in it. mathematica is based on his *own*
> work, not a copycat of someone else's.
Others have responded to this already. I don't know what flow of information
or ideas went between Macsyma/Mathematica (or SMP) but please understand
that I did not mean the term "copycat" in a pergorative sense. And as to his
work being his own ... well, there's very little new under the sun, and
most people's work is to a certain extent a copy of someone else's (and I
contend that that is not necessarily a Bad Thing).
--
Duane Rettig ·····@franz.com Franz Inc. http://www.franz.com/
555 12th St., Suite 1450 http://www.555citycenter.com/
Oakland, Ca. 94607 Phone: (510) 452-2000; Fax: (510) 452-0182
This has already been answered, namely Wolfram used Macsyma
before building SMP. He assumed that since he was very smart
that he could write a program that could do what his Macsyma
program did, but faster, if he wrote it from scratch. He and
a few others at Caltech then decided to build a whole system.
Some of the decisions in SMP were surprisingly bad, like
representing rational numbers as floats. Converting 1/3
to 0.3333333334 is maybe good enough for government work,
but not good enough for mathematicians. I don't think SMP
came out the way he wanted, but in any case the ownership
of SMP was in dispute. The eventual purchaser paid off
both Caltech and Wolfram. The version 1.0 manual I have says
"SMP as described below was developed at California Institute
of Technology, and is the property of California Institute
of Technology." The part of the sentence following the comma
is blacked out with "magic marker". Wolfram later went to
the Institute for Advanced Study, where he apparently made
special arrangments for the disposition of intellectual property,
even going so far as to get a Sun workstation not under
IAS ownership.
Anyway, returning to those days at Caltech in the 1979-81 period...
At his invitation (and maybe on his dollar) I visited Wolfram
at Caltech where he, Chris Cole, Geoff Fox, Jeffrey Greif,
Eric Mjolsness, Larry Romans, Timothy Shaw and Anthony Terrano
[mentioned in the manual] and also perhaps R. Pike, H. Redelmeier..
were working on SMP on the DEC VAX 11/780, a type of computer
which appeared at Berkeley in 1978, and was
running UNIX, Franz Lisp and Macsyma within about 12 months.
The first Caltech VAX arrived in early 1980.
During my visit, Stephen would explain something
about what he was trying to program, and I would say things like
"Oh, this structure you are building is called a Symbol Table."
etc. I believe Wolfram had read volumes 1,2 of Knuth and parts
of the UNIX manuals, but maybe not much else.
As for whether Mathematica is lisp like... well you might think
so because, for example,
Function[{x,y},x+y] [3,4] evaluates to 7
But the dominant paradigm for evaluation is based on pattern matching.
In many cases this looks like ordinary function definition in
other languages, e.g.
g[x_,y_]:=x+y where we ignore, for the moment the _ character.
But in reality g[x_,y_] is a pattern. Other patterns like
g[Sin[x_],3] could be given different replacements. So there
is even an assertion that Mathematica is um, object oriented, with a
g-method for
{Sin[*],3} etc.
There are lots of other features which serve to make a thorough
hash of the semantics of scope. Evaluation looks
sort of fixed-pointy but not always.
As I said, for many users, it doesn't matter. When it does matter,
I believe the business model they use doesn't have much room
for adopting advice from people who know too much about programming
languages, (or numerical computing).
RJF
PS. Could Wolfram have even looked at the lisp source code for
what eventually became DOE Macsyma, before writing SMP? Maybe,
but I would not expect that he did. The version 1 manual says
"The important decision that SMP should be written in the C
language, rather than in LISP, was made at this stage [about 4
months after the start of design in Nov. 1979]. C was favoured
primarily because of the greater flexibility in internal representation
which it afforded, because of its potential portability and because the
task of codingwas expected to be easier."
Duane Rettig wrote:
> ozan s yigit <··@blue.cs.yorku.ca> writes:
>
>
>>Duane Rettig:
>>
>>* Or, copycat apps grow up around it which may not have the pizazz
>>* of the original, but are utilitarian nevertheless and remove the
>>* "killer" label from the original. Examples of each: emacs,
>>* macsyma -> mathematica.
>>
>>this is quite unfair. wolfram had earlier (1983) written his own symbolic
>>mathematics package called SMP.
>>
>
> Actually, according to web searches, he started it even earlier, in 1979.
> But the Macsyma effort had started in 1967, long before he even started
> using computers in 1973 (at age 14).
>
>
>>i don't know if he saw macsyma at caltech
>>or not; we had both packages, and if memory serves, SMP had not a trace
>>of lisp or any part of macsyma in it. mathematica is based on his *own*
>>work, not a copycat of someone else's.
>>
>
> Others have responded to this already. I don't know what flow of information
> or ideas went between Macsyma/Mathematica (or SMP) but please understand
> that I did not mean the term "copycat" in a pergorative sense. And as to his
> work being his own ... well, there's very little new under the sun, and
> most people's work is to a certain extent a copy of someone else's (and I
> contend that that is not necessarily a Bad Thing).
>
>
[fateman's interesting note on wolfram, smp, mathematica and their relation
(or lack of) to macsyma not repeated]
thanks for the historic details. it certainly clarifies some things.
[next time i see hugh, i'll ask him about SMP. i would have thought
he would fix any issues with numeric representations. he knows better]
oz
---
music is the space between the notes -- claude debussy
··@cs.yorku.ca (ozan s. yigit) writes:
> [fateman's interesting note on wolfram, smp, mathematica and their relation
> (or lack of) to macsyma not repeated]
>
> thanks for the historic details. it certainly clarifies some things.
> [next time i see hugh, i'll ask him about SMP. i would have thought
> he would fix any issues with numeric representations. he knows better]
I'm not sure this is precisely the forum in which to log this fact,
but since Fateman is telling historical stories I wanted to add one.
I was in Pasadena at one point, visiting a friend at Caltech, and
popped in to see Wolfram around the time he was gearing up to write
SMP, I think. If I recall, he was 19 at the time. People around me
informed me that though he was very young, or maybe because of it, he
was on track to win a nobel prize of some sort. I myself worked for
the MIT Macsyma group at the time as an undergrad, perhaps my first
senior year, so I think I must have been a year or two older than him.
He told me that Lisp was "inherently" (I'm pretty sure even after all
this time that this was his exact word) 100 times slower than C and
therefore an unsuitable vehicle. I tried to explain to him that this
was implausible. That he could probably construct an argument for 2-5
that he could at least defend in some prima facie way, but that 100
was ridiculous. (This was in the heyday of Maclisp when it had been shown
to trump Fortran's speed, so probably even 2-5 could be refuted, but at least
taking a position in that range would have left him with some defenses in
a debate. He didn't cite anything credible that I recall to back up this
factor of 100 problem.
I tried to explain why and was not clear why a person smart enough to
"maybe win a nobel prize" couldn't entertain a discussion on the
simple set of concepts involved, whether or not schooled in
computation. It was quite frustrating and he seemed impatient.
He in fact did not purport to be adequately competent on the matter of
computation at the time but he pointed to a stack (literally) of books
(I'd say about a foot high) including the Knuth books, the compiler
book with the dragon on it, and a number of other really standard
texts. He then said "I'm going to read these and then I'll know as
much as you." (Again, I'm pretty sure even now that this is pretty
close to an exact quote. But whether it's exact or not, what struck
me was the incredible arrogance of the remark.) The point seemed
debatable, but I didn't bother to debate it. He seemed deadset on his
goal and once he got to the point where he seemed to feel he could use
as a credential books he had not yet read, there seemed to be no
deflecting him.
My real concern, of course, was not that he was using optimized data
structures so much as that he seemed on target to reintroduce
numerical error back into a world that we had worked hard to make
'exact' (Macsyma used bignums from Lisp) or at least 'arbitrarily
exact' (Macsyma had a derived type called 'bigfloat' that was
internally a pair bignums, acting more or less as a ratio but with
lots of other hidden bits to assure that any decimalization had enough
bits to be precise to a given number of digits). Stephen's aim seemed
to be to sacrifice correctness for speed.
He seemed clear on that the error was not a problem for him. I'm not
a domain expert and I can only assume that he did know what he was
doing when he accepted an approximation over an exact value. Probably
HE had a good reason. But I worried for others, and recall telling
him so. My specific recollection of this part of the discussion is
less clear, but I'm pretty sure his response was that what tools
others chose to use or not use was not his concern.
There's a fine ethical line here between simply making a tool and
actively promoting it, but I'll not expound on that in detail.
Rather, I'll just say that this line concerned me. The problem I
have, and had then, is that other users, not him, might NOT understand
that this trade-off had been made and so might not be making an
informed choice. People tend to say "well, if it's good enough for
him" rather than "well, if it's good enough for the purpose he had in
mind" since presumably there are other applications Stephen or anyone
could have come up with that would not have tolerated error.
But ah well. History marched on and what happened happened.
My remarks are just personal memories and opinions, offered just as a
way to add historical perspective for those who care about such
things. I think it's relevant to the Lisp community because it
relates to Stephen's departure from Macsyma (and implicitly from
Lisp)--he had previously been a Macsyma user and I'm pretty sure he
was more annoyed with some specific Macsyma program (that probably
used some inefficient data structures that the underlying Lisp would
have required it to do)... He seemed annoyed at Lisp but I didn't have
the sense that he'd done enough survey of other Lisp programs or of
the language itself to really have a believable opinion on this. Then
again, I'm sure he's not the only person in Lisp history to see one
bad program and then overgeneralize...
[kent's story and opinions of his encounter with wolfram not repeated]
it looks like duane's arrow from macsyma to mathematica *is* meaningful
but different than the way he had intended... :-]
oz
--
I see the eigenvalue in thine eye,
I hear the tender tensor in thy sigh.
Bernoulli would have been content to die
Had he but known such a-squared cos 2(phi)! -- stanislaw lem (cyberiad)
From: Lieven Marchand
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <87of92tlzs.fsf@wyrd.be>
Duane Rettig <·····@franz.com> writes:
> In general, once a killer app gains acceptance, its language is
> forgotten or played down, and the app itself becomes more mundane.
A nice example of the former is cvsup. It's a crucial piece of the
development infrastructure for the *BSD's. It's written in Modula-3
and it would be hard to port it to C or Java. In fact, the author
finds it easier to carry along his own Modula-3 distribution. It
hasn't made much of difference in popularity for Modula-3 though.
--
When they ran out of ships, they used guns. When they ran out of guns,
they used knives and sticks and bare hands. They did this for two
years. They never ran out of courage. But in the end, they ran out of
time. B5 --- In the Beginning
From: Jacek Generowicz
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <tyflm479k7c.fsf@pcitapi22.cern.ch>
Tim Lavoie <········@spamcop.net> writes:
> Lisp needs that app which lots of people will use, at a level where
> they can find it too.
How about a highly configuarable editor? for example :-)
How many people have progressed from tentatively chucking in a single
elisp form into their .emacs, to writing serious elisp code ? And yet
the idea of using elisp to do anything other than enhancing Emacs'
functionality rarely crosses their minds.
Imagine how different this would be if they started off by tentatively
chucking in a single CL form into their .clemacs ...
From: Tim Lavoie
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <87fzuer26h.fsf@theasylum.dyndns.org>
>>>>> "Jacek" == Jacek Generowicz <················@cern.ch> writes:
Jacek> Tim Lavoie <········@spamcop.net> writes:
>> Lisp needs that app which lots of people will use, at a level
>> where they can find it too.
Jacek> How about a highly configuarable editor? for example :-)
Yup, I use that one too.
Jacek> How many people have progressed from tentatively chucking
Jacek> in a single elisp form into their .emacs, to writing
Jacek> serious elisp code ? And yet the idea of using elisp to do
Jacek> anything other than enhancing Emacs' functionality rarely
Jacek> crosses their minds.
Perhaps that is because they don't know yet that they're using the
Emacs Operating System. *I* know that I can read Usenet, send mail,
and show a Towers of Hanoi demo, but most people wouldn't think of
it.
There may be a difference (if only perceived) that they have a text
editor with Lisp embedded in it, rather than a general-purpose
programming language which happens to be used to implement a powerful
text editor. I don't know, can you fire up elisp without the editor? I
haven't tried, but I suspect that it changes what you would think of
doing with it.
Jacek> Imagine how different this would be if they started off by
Jacek> tentatively chucking in a single CL form into their
Jacek> .clemacs ...
:)
--
Those who can't write, write manuals.
Tim Lavoie <········@spamcop.net> writes:
> There may be a difference (if only perceived) that they have a text
> editor with Lisp embedded in it, rather than a general-purpose
> programming language which happens to be used to implement a powerful
> text editor. I don't know, can you fire up elisp without the editor? I
> haven't tried, but I suspect that it changes what you would think of
> doing with it.
It's not worth doing (you'd really want to just use ielm instead), but
you certainly /can/ do it:
$ emacs -batch -eval '(while t (print (eval (read))))'
Lisp expression: (+ 1 1)
2
Lisp expression: (throw 'foo 1)
No catch for tag: foo, 1
$
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
From: Tim Lavoie
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <87wunqroof.fsf@theasylum.dyndns.org>
>>>>> "Thomas" == Thomas F Burdick <···@apocalypse.OCF.Berkeley.EDU> writes:
Thomas> It's not worth doing (you'd really want to just use ielm
Thomas> instead), but
Thomas> you certainly /can/ do it:
Thomas> $ emacs -batch -eval '(while t (print (eval (read))))'
Thomas> Lisp expression: (+ 1 1)
Sure enough, it works. It does seem vaguely disturbing though. :)
--
Things will get better despite our efforts to improve them.
-- Will Rogers
Tim Lavoie <········@spamcop.net> writes:
> >>>>> "Jacek" == Jacek Generowicz <················@cern.ch> writes:
>
> Jacek> How many people have progressed from tentatively chucking
> Jacek> in a single elisp form into their .emacs, to writing
> Jacek> serious elisp code ? And yet the idea of using elisp to do
> Jacek> anything other than enhancing Emacs' functionality rarely
> Jacek> crosses their minds.
>
> Perhaps that is because they don't know yet that they're using the
> Emacs Operating System. *I* know that I can read Usenet, send mail,
> and show a Towers of Hanoi demo, but most people wouldn't think of
> it.
I was alluding to the ones who _are_ aware of all this. There are very
many of them. But for most of them Lisp is still something which is
very much boxed in within emacs, and they wouldn't dream of using it
for "stand alone" applications.
> There may be a difference (if only perceived) that they have a text
> editor with Lisp embedded in it, rather than a general-purpose
> programming language which happens to be used to implement a
> powerful text editor.
Firstly, I think that CL is a much better general purpose programming
language than elisp. Having CL as emacs' scripting language would
increase the chances of users recognizing Lisp as a general purpose
language independent of emacs.
Secondly, parts of emacs are written in C. Having emacs completely
implemented in CL, would probably help significantly to expose Lisp as
a useful general purpose programming language.
Elisp is a scripting language implemented within one application. CL
would be a general purpose language used to implement the actual
application itself, as well as the scripting language _of_ the
application.
I think that this would (pleasantly) shock a lot of people, and
glaringly expose one of Lisp's big advantages.
> I don't know, can you fire up elisp without the editor? I haven't
> tried, but I suspect that it changes what you would think of doing
> with it.
This is what I am talking about: a change in attitude towards the
language which opens a whole new world of possibilities.
>>>>> "Jacek" == Jacek Generowicz <················@cern.ch> writes:
Jacek> I was alluding to the ones who _are_ aware of all
Jacek> this. There are very many of them. But for most of them
Jacek> Lisp is still something which is very much boxed in within
Jacek> emacs, and they wouldn't dream of using it for "stand
Jacek> alone" applications.
Jacek> Firstly, I think that CL is a much better general purpose
Jacek> programming language than elisp. Having CL as emacs'
Jacek> scripting language would increase the chances of users
Jacek> recognizing Lisp as a general purpose language independent
Jacek> of emacs.
Jacek> Secondly, parts of emacs are written in C. Having emacs
Jacek> completely implemented in CL, would probably help
Jacek> significantly to expose Lisp as
Jacek> a useful general purpose programming language.
Jacek> Elisp is a scripting language implemented within one
Jacek> application. CL would be a general purpose language used to
Jacek> implement the actual application itself, as well as the
Jacek> scripting language _of_ the application.
Jacek> I think that this would (pleasantly) shock a lot of people,
Jacek> and glaringly expose one of Lisp's big advantages.
>> I don't know, can you fire up elisp without the editor? I
>> haven't tried, but I suspect that it changes what you would
>> think of doing with it.
Jacek> This is what I am talking about: a change in attitude
Jacek> towards the language which opens a whole new world of
Jacek> possibilities.
The "Emacs in CL" thread has come up before, hasn't it? I agree of
course, but I suspect that the reasons it hasn't happened are more
political than technical. I'm no Lisp expert, so I can't exactly
volunteer to just jump in and do it. It could be sweet though.
Tim
--
"Why does man kill? He kills for food. And not only food: frequently
there must be a beverage."
--Woody Allen
The world rejoiced as Tim Lavoie <········@spamcop.net> wrote:
> The "Emacs in CL" thread has come up before, hasn't it? I agree of
> course, but I suspect that the reasons it hasn't happened are more
> political than technical. I'm no Lisp expert, so I can't exactly
> volunteer to just jump in and do it. It could be sweet though.
There's certainly a combination of both political and technical
considerations involved. And some of the "political" aspects
/necessitate/ technical choices.
For instance, picking a CL implementation is a thorny matter:
- Picking ACL/LispWorks buys some platform independence, but
requires that users get ACL/LispWorks licenses. The Emacs
code may be "free," but the aggregate environment isn't.
- The notable "libre" CLs are somewhat unsatisfactory in
varying ways, /technically/.
Erik Naggum had a proposal that amounted to compiling CL into portable
C, which would get around this; that is by no means a trivial thing to
do.
Picking a CL implementation to use (or ensuring the result is
"portable" across several) is, all by itself, a big /political/ matter
that is a mandatory prerequisite to writing the first line of code.
The only way I'd think there is for the "politics to be irrelevant in
a perfect world" would be for RMS' "world without copyright" to get
implemented, and it's fair to say that not everyone finds that
'utopia' to be their cup of tea.
--
(reverse (concatenate 'string ·············@" "enworbbc"))
http://cbbrowne.com/info/sgml.html
"I don't plan to maintain it, just to install it." -- Richard M. Stallman
From: Erik Naggum
Subject: Re: CLEmacs [was: I'm a Lisper, hear me roar...]
Date:
Message-ID: <3245707024453408@naggum.no>
* Tim Lavoie
| The "Emacs in CL" thread has come up before, hasn't it? I agree of
| course, but I suspect that the reasons it hasn't happened are more
| political than technical. I'm no Lisp expert, so I can't exactly
| volunteer to just jump in and do it. It could be sweet though.
An Emacs in Common Lisp would need a full reimplementation of Emacs Lisp
and would need an enormous compatibility layer if it were to do anything
differently internally. To succeed, it would also need to emulate both
XEmacs and Emacs. Then it would need to track the development of both
Emacsen. While it would make a lot of sense to reimplement the internals
so that the MULEshit could be cleaned out and international support done
right, Emacs is not just internals. What would /really/ make a difference
to users would be if Emacs Lisp could be compiled to native code and not
have to run through the byte-code interpreter.
Some of the other things I have wanted for a Common Lisp Emacs is a user
process that handles file system interaction instead of the Emacs doing
it directly, a separation of the displayable area and the "buffer" so
that less of the data would need to be transferred to the Emacs process
and converted to the internal character set, which should be Unicode, and
such that changes to the actual file would require minimal work, and many
other things that are very hard to do in Emacs today that would change
the way Emacs behaves and which would make it so much more usable for
large files and being a "control center" now that most computer users
command a number of computer, usually widely dispersed. Some of these
ideas are strong enough that they could make a Common Lisp Emacs survive,
but the amount of work to get there would simply be enormous.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
>>>>> "Erik" == Erik Naggum <····@naggum.no> writes:
Erik> Some of these ideas are strong enough that they could make a
Erik> Common Lisp Emacs survive, but the amount of work to get
Erik> there would simply be enormous.
Well, I guess that sums it up, thanks for the insights. The sheer
volume of add-on code for existing Emacs versions makes the idea more
attractive, and the process more daunting.
Cheers,
Tim
Tim Lavoie <········@spamcop.net> writes:
> The "Emacs in CL" thread has come up before, hasn't it?
Oh yes.
Have the thoughts on this been distilled and localized somewhere?
Googiling around the subject in the past, I couldn't find the signal
amongst the noise within a sensible amount of time. There are links to
pages by Erik, but these all seemed to be dead.
Erik Naggum <····@naggum.no> writes:
> the amount of work to get there would simply be enormous.
Presumably the Guile movement will be faced with similar problems, to
some extent. Have they got a plan, or is it just hoping that wherever RMS
takes them, hordes of developers will follow?
Hmm ... A quick look suggests that the plan is merely to stick guile
onto Emacs. Now I remember why I was so completely underwhelemed by
the whole idea.
From: Erik Naggum
Subject: Re: CLEmacs [was: I'm a Lisper, hear me roar...]
Date:
Message-ID: <3245729292637028@naggum.no>
* Jacek Generowicz
| Have the thoughts on this been distilled and localized somewhere?
Well, how about the destillation: "It turned out to be a waste of time."
| Hmm ... A quick look suggests that the plan is merely to stick guile
| onto Emacs. Now I remember why I was so completely underwhelemed by the
| whole idea.
I fear the whole Guile thing may be the end of Emacs as we know it.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Erik Naggum <····@naggum.no> writes:
> * Jacek Generowicz
> | Have the thoughts on this been distilled and localized somewhere?
>
> Well, how about the destillation: "It turned out to be a waste of time."
Certainly concise. Possibly accurate. However, it doesn't help me to
improve my understanding of the issues involved, the possible
approaches, or the reasons why it is a waste of time.
Erik Naggum <····@naggum.no> writes:
> * Tim Lavoie
> | The "Emacs in CL" thread has come up before, hasn't it? I agree of
> | course, but I suspect that the reasons it hasn't happened are more
> | political than technical. I'm no Lisp expert, so I can't exactly
> | volunteer to just jump in and do it. It could be sweet though.
>
> An Emacs in Common Lisp would need a full reimplementation of
> Emacs Lisp and would need an enormous compatibility layer if it
> were to do anything differently internally. To succeed, it would
> also need to emulate both XEmacs and Emacs. Then it would need to
> track the development of both Emacsen.
Serious question: Why?
Surely there's a case for throwing out the 'pure' backward
compatibility (as happened when emacs was ported from TECO macros to
Lisp) in favour of getting something working with a sane set of
internals and a clean programmatic interface to them. Initially one
could do ports of 'important' elisp modules to the new language by
hand, bearing in mind that it's generally better to port by writing a
pile of support functions rather than by changing the code itself.
It's still a *huge* pile of work of course, but by doing it
incrementally and worrying about the elisp interface later, you at
least get yourself a working, useful editor a good deal earlier in
your development cycle.
> While it would make a lot of sense to reimplement the internals so
> that the MULEshit could be cleaned out and international support
> done right, Emacs is not just internals. What would /really/ make
> a difference to users would be if Emacs Lisp could be compiled to
> native code and not have to run through the byte-code interpreter.
No question, but again, I'm not sure that it's something that's
absolutely necessary for the initial releases.
Not that I actually have the skills to put my money where my mouth is
here. Feel free to dismiss me as a crank, I was just curious as to
whether this sort of incremental approach had been considered.
--
Piers
"It is a truth universally acknowledged that a language in
possession of a rich syntax must be in need of a rewrite."
-- Jane Austen?
Piers Cawley <········@bofh.org.uk> writes:
> Surely there's a case for throwing out the 'pure' backward
> compatibility (as happened when emacs was ported from TECO macros to
> Lisp) in favour of getting something working with a sane set of
> internals and a clean programmatic interface to them.
this has been done several times. The best example is FRED (Fred
Resembles Emacs Deliberately), the editor of Macintosh Common Lisp,
which is has a nice CLOS API.
--
(espen)
Espen Vestre <·····@*do-not-spam-me*.vestre.net> writes:
> Piers Cawley <········@bofh.org.uk> writes:
>
> > Surely there's a case for throwing out the 'pure' backward
> > compatibility (as happened when emacs was ported from TECO macros to
> > Lisp) in favour of getting something working with a sane set of
> > internals and a clean programmatic interface to them.
>
> this has been done several times. The best example is FRED (Fred
> Resembles Emacs Deliberately), the editor of Macintosh Common Lisp,
> which is has a nice CLOS API.
I smell chicken and egg here :) MCL alse has a nice CLOS GUI.
Cheers
--
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group tel. +1 - 212 - 998 3488
715 Broadway 10th Floor fax +1 - 212 - 995 4122
New York, NY 10003, USA http://bioinformatics.cat.nyu.edu
"Hello New York! We'll do what we can!"
Bill Murray in `Ghostbusters'.
* Piers Cawley <········@bofh.org.uk>
| Serious question: Why?
Because even now, package developers are seriously pissed off by the
incompatibilities between Emacs and XEmacs.
| Initially one could do ports of 'important' elisp modules to the new
| language by hand, bearing in mind that it's generally better to port by
| writing a pile of support functions rather than by changing the code
| itself.
Emacs Lisp is the extension language built on top of the C substrate,
which exports a large number of C-level decisions to Emacs Lisp. In
order for such a project to get off the ground at all, those packages
that people are quite used to would have to work in the new Emacs.
This means presenting a compatibility layer that is quite extensive.
Take my word for it, if you want to change the substrate language, you
will notice that the substrate language of Emacs really is C. This is
not something people will believe right away, since it superficially
looks very much like a Lisp application.
| It's still a *huge* pile of work of course, but by doing it incrementally
| and worrying about the elisp interface later, you at least get yourself a
| working, useful editor a good deal earlier in your development cycle.
Well, I went there, spent a few months hacking on clemacs (and even got
the domain name), and discovered that it would take many years, for no
real practical improvements until /way/ into the project.
| Not that I actually have the skills to put my money where my mouth is
| here. Feel free to dismiss me as a crank, I was just curious as to
| whether this sort of incremental approach had been considered.
All approaches you can think have been considered...
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Piers Cawley <········@bofh.org.uk> writes:
> Erik Naggum <····@naggum.no> writes:
>
> > * Tim Lavoie
> > | The "Emacs in CL" thread has come up before, hasn't it? I agree of
> > | course, but I suspect that the reasons it hasn't happened are more
> > | political than technical. I'm no Lisp expert, so I can't exactly
> > | volunteer to just jump in and do it. It could be sweet though.
> >
> > An Emacs in Common Lisp would need a full reimplementation of
> > Emacs Lisp and would need an enormous compatibility layer if it
> > were to do anything differently internally. To succeed, it would
> > also need to emulate both XEmacs and Emacs. Then it would need to
> > track the development of both Emacsen.
>
> Serious question: Why?
Because without the elisp packages it is not XEmacs/Emacs. A good
example of this is the "emacs like" editor that uses Ocaml as the user
language, if I remember correctly it is written in Ocaml also. ^x^c
does not make emacs, all the elisp packages that have been developed
over many years make it emacs. The reimplementation effort would be
/huge/. From what little I know we would be talking about junking
hundreds, possibly thousands, of man/years of work.
Also think of the political effort needed, you would be declaring
effective war on the FSF's vision of emacs(guile). From what I have
read the Emacs/XEmacs split was a very ugly affair.
Also to do this you need a number of good programmers to spend
a serious chunk of time, each, to make it happen. How do they get
compensated for the work? The simple face is that if people are not
willing to pay for it, it is not important. Money is honest.
>
> > While it would make a lot of sense to reimplement the internals so
> > that the MULEshit could be cleaned out and international support
> > done right, Emacs is not just internals. What would /really/ make
> > a difference to users would be if Emacs Lisp could be compiled to
> > native code and not have to run through the byte-code interpreter.
>
> No question, but again, I'm not sure that it's something that's
> absolutely necessary for the initial releases.
It would not be emacs until it ran the elisp packages, IMO. You would
just have an editor written in common lisp, not a bad thing but not
emacs.
marc
* Piers Cawley wrote:
> It's still a *huge* pile of work of course, but by doing it
> incrementally and worrying about the elisp interface later, you at
> least get yourself a working, useful editor a good deal earlier in
> your development cycle.
Well, what happens is that you end up with two editors, which are
incompatible.
If we imagine a lone developer doing this (I think it scales to lots
of people), then you find that this person uses a huge number of
random Emacs packages all the time, and that they do things that are
useful to them. Little tools to maintain file headers, do cleverer
line wrapping, incrementally spell-check text files and mail messages,
make writing TeX pleasant, make writing HTML bearable, process files
in outline mode, edit lisp/C/Fortran/Java comfortably, insert
`signatures', integrate with 4 different source-control systems, read
news, read mail, &c &c. Hundreds of the things.
The moment you get your CL-based editor off the ground, about the
first thing you find is that *none of this works*. Even if the editor
is highly compatible with Emacs in it's `way of being' so things like
keystrokes work as you expect, nothing else works. You still have to
use Emacs all the time. Even if you have one thing that this editor
does better than Emacs - say it integrates with a CL system incredibly
well - then you still need two editors, because there is so much stuff
that Emacs does that you find you really need. Until you've either
made this thing so you can simply use elisp, or reimplemented all the
Emacs stuff you use, you're stuck in this awkward position.
I went through this with hemlock. Hemlock had Lisp interaction to die
for: at the time ilisp was a bad joke which would regularly crash
emacs, and even when it didn't hardly worked, but hemlock's
integration with CMUCL was a delight. I even wrote some clever
form-selection stuff in Hemlock, based on what TEdit did, which would
have been significantly more work in Emacs (then). But that was *all*
Hemlock did - it wouldn't run news, talk to CVS, keep file headers up
to date &c &c. And every 3 days some new cool package would appear
for Emacs, but not for Hemlock. Eventually I stopped using it, when
lemacs started working and supported X properly, unlike emacs 18.
--tim
Tim Bradshaw <···@cley.com> writes:
> * Piers Cawley wrote:
> > It's still a *huge* pile of work of course, but by doing it
> > incrementally and worrying about the elisp interface later, you at
> > least get yourself a working, useful editor a good deal earlier in
> > your development cycle.
>
> Well, what happens is that you end up with two editors, which are
> incompatible.
>
> If we imagine a lone developer doing this (I think it scales to lots
> of people), then you find that this person uses a huge number of
> random Emacs packages all the time, and that they do things that are
> useful to them. Little tools to maintain file headers, do cleverer
...
>
> The moment you get your CL-based editor off the ground, about the
> first thing you find is that *none of this works*. Even if the editor
I'm not sure why any of this matters. Is Emacs really the only tool
you use? Would you disregard a newreader out of hand if it didn't
integrate with emacs?
I think that if a cl-emacs "takes off" it will be on its own merits,
not as a GNU Emacs clone. I'm not sure what the driving force will
be; perhaps good integration with the development environment and CLIM,
perhaps real editing with shared kill ring in every text widget.
> I went through this with hemlock. Hemlock had Lisp interaction to die
> for: at the time ilisp was a bad joke which would regularly crash
> emacs, and even when it didn't hardly worked, but hemlock's
> integration with CMUCL was a delight. I even wrote some clever
> form-selection stuff in Hemlock, based on what TEdit did, which would
> have been significantly more work in Emacs (then). But that was *all*
> Hemlock did - it wouldn't run news, talk to CVS, keep file headers up
> to date &c &c. And every 3 days some new cool package would appear
> for Emacs, but not for Hemlock. Eventually I stopped using it, when
> lemacs started working and supported X properly, unlike emacs 18.
This doesn't really explain why you stopped using Hemlock... was
unbearable to switch between editors for different tasks?
Tim
+ Timothy Moore <·····@pac-dhcp-22-84.amazon.com>:
| This doesn't really explain why you stopped using Hemlock... was
| unbearable to switch between editors for different tasks?
I don't know why he did, but to me it's a royal pain switching between
programs having similar but not quite the same keybindings. Of
course, any program worth its salt lets you change keybindings (except
vi - I'm not sure about vim), but unless I use it a lot I don't
usually bother. I mean, how much time do you want to spend diddling
with your environment instead of doing real work? I guess I am really
conservative on this front. I've only ever used three window managers
(twm, fvwm, sawfish) and don't really want to change again unless
bitrot makes the current one totally useless. But I seem to have
gotten off along a tangent here... Even if you can change keybindings
to make to editors more similar, there will always be differences that
tend to confuse, and paradoxically the more subtle the differences,
the more likely they are to trap you. When Norwegians and Swedes
converse, misunderstandings often occur because we think we understand
each other's language perfectly well, when in fact we do not. So in
summary, I am much more comfortable switching between emacs and vi
than between emacs and some other emacs-like editor. In fact, I often
prefer to use vim for many trivial editing jobs.
--
* Harald Hanche-Olsen <URL:http://www.math.ntnu.no/~hanche/>
- Yes it works in practice - but does it work in theory?
"Harald Hanche-Olsen" <······@math.ntnu.no> wrote in message
····················@thoth.math.ntnu.no...
> + Timothy Moore <·····@pac-dhcp-22-84.amazon.com>:
>
> | This doesn't really explain why you stopped using Hemlock... was
> | unbearable to switch between editors for different tasks?
>
> I don't know why he did, but to me it's a royal pain switching between
> programs having similar but not quite the same keybindings.
This is quite a valid concern. It is fairly easy to switch from vi to emacs
simply because they are so dramatically different in the key bindings and
how they are used. However, for me, when I tried to switch from GNU Emacs to
XEmacs on Windows, I simply gave up. XEmacs bindings are Different. Some
could say "better" because they're closer to other "standards". Be that as
it may, they were NOT the bindings that I was used to, and I'm not enough of
an Emacs type of person where rebinding the keys to the way I like would
have been a trivial process. It simply wasn't worth my time. And I wasn't
really motivated to relearn the bindings either.
So, I stick with the GNU Emacs. It's the old Baby Duck syndrome.
However I do agree with Tim there's no reason why CLEmacs needs to "Do
Everything" on day one, and that perhaps it could just provide a better
environment and tighter integration with a CL image.
Regards,
Will Hartung
(·····@msoft.com)
>>>>> On Sat, 09 Nov 2002 16:49:31 GMT, Will Hartung ("Will") writes:
Will> However, for me, when I tried to switch from GNU Emacs to
Will> XEmacs on Windows, I simply gave up. XEmacs bindings are Different.
I've been using XEmacs and GNU Emacs interchangably for
the last few weeks, and didn't notice anything different
(except that it has fonts).
* Christopher C. Stacy
| I've been using XEmacs and GNU Emacs interchangably for
| the last few weeks, and didn't notice anything different
| (except that it has fonts).
Huh? I read this with GNU Emacs 21.2 and Gnus 5.9 under X11 and the font
is Adobe Utopia, a Unicode-encoded proportional font while the normal
editing font is Lucida Typewriter. This would have been impossible with
GNU Emacs 20, but it is now common on Emacs 21 to use different fonts.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Now showing on CNN: Harry Potter and the Search for Weapons of Mass Destruction
>>>>> On 09 Nov 2002 19:53:12 +0000, Erik Naggum ("Erik") writes:
Erik> * Christopher C. Stacy
Erik> | I've been using XEmacs and GNU Emacs interchangably for
Erik> | the last few weeks, and didn't notice anything different
Erik> | (except that it has fonts).
Erik> Huh? I read this with GNU Emacs 21.2 and Gnus 5.9 under X11 and the font
Erik> is Adobe Utopia, a Unicode-encoded proportional font while the normal
Erik> editing font is Lucida Typewriter. This would have been impossible with
Erik> GNU Emacs 20, but it is now common on Emacs 21 to use different fonts.
I run GNU Emacs 21.2 also, but I don't get fonts in Gnus.
But I'm running under MS Windows, so maybe it's different under X.
In any event, your comment would seem to support my view that
GNU Emacs and XEmacs are highly compatible.
My point in particular was to refute the claim that many
keybindings between the two Emacsen are different.
A few weeks ago I downloaded XEmacs just to see what it was like.
I remember not liking it a long time ago (mainly because of the
floating minibuffer window, which is no longer the default).
My experience is: the key bindings for everything I've tried so far
are all the same, and I haven't had to adjust or learn anything, and I
switch back and forth between Emacs and XEmacs whenever I feel like it.
(And I've been using Emacs exclusively as my editor for over 22 years,
so it's not like I'm a naive user who doesn't exercise the features or
something.)
Also, all my customizations (including several libraries I wrote)
all worked. The only thing I had to change was the syntax of some
keystrokes: you write [C-insert] as [(control insert)] in Xemacs.
At some point I'll probably decide to pick one or the other.
XEmacs is much easier to install and comes preconfigured with lots
of packages, so that's what I am now currently recommending to my
friends who are Windows users but who have not tried Emacs before.
Centuries ago, Nostradamus foresaw when ······@dtpq.com (Christopher C. Stacy) would write:
>>>>>> On Sat, 09 Nov 2002 16:49:31 GMT, Will Hartung ("Will") writes:
> Will> However, for me, when I tried to switch from GNU Emacs to
> Will> XEmacs on Windows, I simply gave up. XEmacs bindings are Different.
>
> I've been using XEmacs and GNU Emacs interchangably for the last few
> weeks, and didn't notice anything different (except that it has
> fonts).
I'm not sure how to explain it, but when I started a new contract
recently, GNU Emacs was what got installed initially, and within days
I found it too painful /not/ to throw in XEmacs.
That's not to "bash" GNU Emacs as being horrible, just that I have
used XEmacs so much the last year or two that there's a significant
impedance of "feel" between them. I'm quite sure that if I had been
using GNU Emacs, the exact opposite would be true.
--
(reverse (concatenate 'string ········@" "enworbbc"))
http://cbbrowne.com/info/oses.html
The nice thing about Windows is - It does not just crash, it displays
a dialog box and lets you press 'OK' first. (Arno Schaefer's .sig)
······@dtpq.com (Christopher C. Stacy) writes:
> >>>>> On Sat, 09 Nov 2002 16:49:31 GMT, Will Hartung ("Will") writes:
> Will> However, for me, when I tried to switch from GNU Emacs to
> Will> XEmacs on Windows, I simply gave up. XEmacs bindings are Different.
>
> I've been using XEmacs and GNU Emacs interchangably for
> the last few weeks, and didn't notice anything different
> (except that it has fonts).
beginning-of-buffer and end-of-buffer behave subtly different.
In GNU Emacs, they set point, in XEmacs they don't.
This has bitten me multiple times.
//Ingvar
--
When in douFNORD! This signature has been hi-jacked by Fnord Information
systems, to fnordprovide you with unfnordlimited information.
>>>>> On 11 Nov 2002 17:56:39 +0000, Ingvar Mattsson ("Ingvar") writes:
Ingvar> ······@dtpq.com (Christopher C. Stacy) writes:
>> >>>>> On Sat, 09 Nov 2002 16:49:31 GMT, Will Hartung ("Will") writes:
Will> However, for me, when I tried to switch from GNU Emacs to
Will> XEmacs on Windows, I simply gave up. XEmacs bindings are Different.
>>
>> I've been using XEmacs and GNU Emacs interchangably for
>> the last few weeks, and didn't notice anything different
>> (except that it has fonts).
Ingvar> beginning-of-buffer and end-of-buffer behave subtly different.
Ingvar> In GNU Emacs, they set point, in XEmacs they don't.
Ingvar> This has bitten me multiple times.
"Point" is where the cursor is, so of course they move the point.
You must have meant to say "Mark". But you are incorrect anyway:
those commands do set the mark. I tried it just now.
I'm using XEmacs 21.4.10, and GNU Emacs 21.2.1.
······@dtpq.com (Christopher C. Stacy) writes:
> >>>>> On 11 Nov 2002 17:56:39 +0000, Ingvar Mattsson ("Ingvar") writes:
>
> Ingvar> ······@dtpq.com (Christopher C. Stacy) writes:
> >> >>>>> On Sat, 09 Nov 2002 16:49:31 GMT, Will Hartung ("Will") writes:
> Will> However, for me, when I tried to switch from GNU Emacs to
> Will> XEmacs on Windows, I simply gave up. XEmacs bindings are Different.
> >>
> >> I've been using XEmacs and GNU Emacs interchangably for
> >> the last few weeks, and didn't notice anything different
> >> (except that it has fonts).
>
> Ingvar> beginning-of-buffer and end-of-buffer behave subtly different.
>
> Ingvar> In GNU Emacs, they set point, in XEmacs they don't.
>
> Ingvar> This has bitten me multiple times.
>
> "Point" is where the cursor is, so of course they move the point.
> You must have meant to say "Mark". But you are incorrect anyway:
> those commands do set the mark. I tried it just now.
Hrm, yes, I meant "sets mark". I was informed by an XEmacs user that
said behaviour was still present in whatever version he was using as
of "a few weeks ago".
> I'm using XEmacs 21.4.10, and GNU Emacs 21.2.1.
Interestingly enough, XEmacs has, indeed, switched.
XEmacs as of fairly current two years ago did *not* behave taht way,
to my extreme irritation. I wonder when the switch happened?
//Ingvar
--
Self-referencing
Five, seven, five syllables
This haiku contains
>>>>> On 12 Nov 2002 14:50:24 +0000, Ingvar Mattsson ("Ingvar") writes:
Ingvar> ······@dtpq.com (Christopher C. Stacy) writes:
>> >>>>> On 11 Nov 2002 17:56:39 +0000, Ingvar Mattsson ("Ingvar") writes:
>>
Ingvar> ······@dtpq.com (Christopher C. Stacy) writes:
>> >> >>>>> On Sat, 09 Nov 2002 16:49:31 GMT, Will Hartung ("Will") writes:
Will> However, for me, when I tried to switch from GNU Emacs to
Will> XEmacs on Windows, I simply gave up. XEmacs bindings are Different.
>> >>
>> >> I've been using XEmacs and GNU Emacs interchangably for
>> >> the last few weeks, and didn't notice anything different
>> >> (except that it has fonts).
>>
Ingvar> beginning-of-buffer and end-of-buffer behave subtly different.
>>
Ingvar> In GNU Emacs, they set point, in XEmacs they don't.
>>
Ingvar> This has bitten me multiple times.
>>
>> "Point" is where the cursor is, so of course they move the point.
>> You must have meant to say "Mark". But you are incorrect anyway:
>> those commands do set the mark. I tried it just now.
Ingvar> Hrm, yes, I meant "sets mark". I was informed by an XEmacs user that
Ingvar> said behaviour was still present in whatever version he was using as
Ingvar> of "a few weeks ago".
>> I'm using XEmacs 21.4.10, and GNU Emacs 21.2.1.
Ingvar> Interestingly enough, XEmacs has, indeed, switched.
Ingvar> XEmacs as of fairly current two years ago did *not* behave taht way,
Ingvar> to my extreme irritation. I wonder when the switch happened?
I bet that it did not change and that your friend was mistaken,
* Ingvar Mattsson wrote:
> Hrm, yes, I meant "sets mark". I was informed by an XEmacs user that
> said behaviour was still present in whatever version he was using as
> of "a few weeks ago".
I bet that the difference is that while both versions set the mark,
xemacs does not make the region active, so many commands which work on
the region will get an error. The change may be that the default
value of zmacs-regions has changed from nil to t.
--tim
>>>>> On 15 Nov 2002 10:06:37 +0000, Tim Bradshaw ("Tim") writes:
Tim> * Ingvar Mattsson wrote:
>> Hrm, yes, I meant "sets mark". I was informed by an XEmacs user that
>> said behaviour was still present in whatever version he was using as
>> of "a few weeks ago".
Tim> I bet that the difference is that while both versions set the mark,
Tim> xemacs does not make the region active, so many commands which work on
Tim> the region will get an error. The change may be that the default
Tim> value of zmacs-regions has changed from nil to t.
The "region" is defined to be the text between "point" and "mark".
The variable "zmacs-regions" is a feature that exists only in XEmacs,
and it emulates the way that Zmacs (the Lisp Machine version of Emacs)
behaved about regions. In Zmacs, the region only existed when it was
explicitly activated as part of some editing commands. Just setting
mark was not enough. Subsequent commands other than motion would
deactivate the region. The purpose of all this was to avoid surprises
from region-hacking commands when you had not realized you had set mark.
This feature is on by default, but you can turn it off by setting
the variable "zmacs-regions".
I don't think they changed the default for this variable, because
turning it ON would be what would cause the "problem". But the
default setting still appears to be "t". If someone was having
trouble with this, they should just do (setq zmacs-regions nil)
in their init file.
But the command that Ingvar was complaining about was m-<
(aka beginning-of-buffer), would not be affected by this anyway
because it activates the region.
I didn't actually know about this feature until you mentioned it.
As I said, I switch back and forth between the two versions of Emacs,
and have never noticed the difference. (Of course, I used to switch
back and forth from GNU Emacs and TECO Emacs and LispM Zmacs, too.)
* Christopher C Stacy wrote:
> The variable "zmacs-regions" is a feature that exists only in XEmacs,
> and it emulates the way that Zmacs (the Lisp Machine version of Emacs)
> behaved about regions. In Zmacs, the region only existed when it was
> explicitly activated as part of some editing commands. Just setting
> mark was not enough. Subsequent commands other than motion would
> deactivate the region. The purpose of all this was to avoid surprises
> from region-hacking commands when you had not realized you had set mark.
> This feature is on by default, but you can turn it off by setting
> the variable "zmacs-regions".
yes, I know this, I've used xemacs since long before it was xemacs
(:-).
> But the command that Ingvar was complaining about was m-<
> (aka beginning-of-buffer), would not be affected by this anyway
> because it activates the region.
No, it doesn't (or it doesn't in my xemacs - 21.1.14 I think). What
it does is not *deactivate* the region. If there isn't already an
active region it won't activate one. I really hope they haven't
changed this in some later version.
--tim
······@dtpq.com (Christopher C. Stacy) writes:
> The variable "zmacs-regions" is a feature that exists only in XEmacs,
> and it emulates the way that Zmacs (the Lisp Machine version of Emacs)
> behaved about regions. [description removed]
In Emacs, it's called Transient Mark Mode, and very useful it is,
especially for a fast typist. It also highlights the region (as on
Zmacs), which I find reassuring. Do (transient-mark-mode 1).
--
Pekka P. Pirinen
In cyberspace, everybody can hear you scream. - Gary Lewandowski
Harald Hanche-Olsen <······@math.ntnu.no> writes:
> + Timothy Moore <·····@pac-dhcp-22-84.amazon.com>:
>
> | This doesn't really explain why you stopped using Hemlock... was
> | unbearable to switch between editors for different tasks?
>
> I don't know why he did, but to me it's a royal pain switching between
> programs having similar but not quite the same keybindings.
You could still use the editor's keybindings for editing text, you
just wouldn't be running, e.g., gnus, inside the editor. This is a
loaded thread to jump into (though most are :)), but I think the
"right" clemacs wouldn't be "emacs with Common Lisp instead of emacs
lisp", I think it would be a programmable text editor that would be
part of a larger GUI framework. You can have a consistent interface
across multiple programs--look at Genera for a superb example of this.
There are certainly some things that definitely belong in a text
editor--things having to do with editing text. This is admittedly a
lot of stuff: a tag system and associated commands, syntax
highlighting, indentation, etc. for all supported languages. But it's
a lot less stuff then re-implementing or providing an emulation mode
for all emacs programs that aren't really about editing text (w3,
gnus, calc, ...).
If I had a clemacs that could edit human languages, I would rather
have it be a pane in my news program that I used for editing messages
then have it be used for all the other functions of the newsreader.
This isn't intended to be a slam on emacs: I'm writing this in gnus
and use emacs an awful lot. It's just confusing to call emacs an
editor when it's really more of a programming environment for emacs
lisp--one that has boxed itself in to being an editor. If you want
clemacs to be just an editor, make it an editor and ignore stuff like
w3, gnus, the mail modes, etc. If you want it to be an
environment--like emacs--make the newsreader, the browser, the mail
reader, etc. separate processes that use the editor but aren't a part
of it. Hopefully McCLIM will be released soon and its editor can be
grown into this role...
Gabe Garza
* Gabe Garza
| This is a loaded thread to jump into (though most are :)), but I think
| the "right" clemacs wouldn't be "emacs with Common Lisp instead of emacs
| lisp", I think it would be a programmable text editor that would be part
| of a larger GUI framework.
But once you had Common Lisp, would anyone want to implement an Emacs
Lisp engine or compiler or byte-code jit compiler or whatever because
they missed just about everything from GNU Emacs? I think so. If that
turned out to be painful, I think they would leave.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Now showing on CNN: Harry Potter and the Search for Weapons of Mass Destruction
Gabe Garza <·······@ix.netcom.com> writes:
> Harald Hanche-Olsen <······@math.ntnu.no> writes:
>
> > + Timothy Moore <·····@pac-dhcp-22-84.amazon.com>:
> >
> > | This doesn't really explain why you stopped using Hemlock... was
> > | unbearable to switch between editors for different tasks?
> >
> > I don't know why he did, but to me it's a royal pain switching between
> > programs having similar but not quite the same keybindings.
>
> You could still use the editor's keybindings for editing text, you
> just wouldn't be running, e.g., gnus, inside the editor. This is a
> loaded thread to jump into (though most are :)), but I think the
> "right" clemacs wouldn't be "emacs with Common Lisp instead of emacs
> lisp", I think it would be a programmable text editor that would be
> part of a larger GUI framework. You can have a consistent interface
> across multiple programs--look at Genera for a superb example of this.
Yes. This was my motivation for writing an editor substrate for
McCLIM -- Goatee -- that can eventually support a real editor, rather
than doing the simplest possible thing that would support command line
editing.
> This isn't intended to be a slam on emacs: I'm writing this in gnus
> and use emacs an awful lot. It's just confusing to call emacs an
> editor when it's really more of a programming environment for emacs
> lisp--one that has boxed itself in to being an editor. If you want
> clemacs to be just an editor, make it an editor and ignore stuff like
> w3, gnus, the mail modes, etc. If you want it to be an
> environment--like emacs--make the newsreader, the browser, the mail
> reader, etc. separate processes that use the editor but aren't a part
> of it. Hopefully McCLIM will be released soon and its editor can be
> grown into this role...
Amen, brother.
Tim
Gabe Garza <·······@ix.netcom.com> writes:
> If I had a clemacs that could edit human languages, I would rather
> have it be a pane in my news program that I used for editing messages
> then have it be used for all the other functions of the newsreader.
How about having every editable text field be an Emacs window?
In article <··············@junk.nocrew.org>,
Lars Brinkhoff <·········@nocrew.org> wrote:
>Gabe Garza <·······@ix.netcom.com> writes:
>> If I had a clemacs that could edit human languages, I would rather
>> have it be a pane in my news program that I used for editing messages
>> then have it be used for all the other functions of the newsreader.
>
>How about having every editable text field be an Emacs window?
That's basically the way MCL handles it -- with a few exceptions for
weight, if you can type in it, it's a fred window. Works fairly nicely.
* Timothy Moore
| Is Emacs really the only tool you use?
Why, yes, of course. (And I'm dead serious, too.) (Well, almost.)
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
Now showing on CNN: Harry Potter and the Search for Weapons of Mass Destruction
Erik Naggum <····@naggum.no> writes:
> * Timothy Moore
> | Is Emacs really the only tool you use?
>
> Why, yes, of course. (And I'm dead serious, too.) (Well, almost.)
>
I hear you :) 5 years ago I would have been right with you, and I am
writing this in Emacs and Gnus. But for me The Web and buying a Mac
have changed things a bit.
Tim
In an attempt to throw the authorities off his trail, Timothy Moore <·····@pac-dhcp-22-84.amazon.com> transmitted:
> Tim Bradshaw <···@cley.com> writes:
>
>> * Piers Cawley wrote:
>> > It's still a *huge* pile of work of course, but by doing it
>> > incrementally and worrying about the elisp interface later, you at
>> > least get yourself a working, useful editor a good deal earlier in
>> > your development cycle.
>>
>> Well, what happens is that you end up with two editors, which are
>> incompatible.
>>
>> If we imagine a lone developer doing this (I think it scales to lots
>> of people), then you find that this person uses a huge number of
>> random Emacs packages all the time, and that they do things that are
>> useful to them. Little tools to maintain file headers, do cleverer
> ....
>>
>> The moment you get your CL-based editor off the ground, about the
>> first thing you find is that *none of this works*. Even if the editor
>
> I'm not sure why any of this matters. Is Emacs really the only tool
> you use? Would you disregard a newreader out of hand if it didn't
> integrate with emacs?
As far as /I/ am concerned, the vital things that /I/ want are the
ability to run Gnus more efficiently, to edit SGML documents more
efficiently, and to otherwise be able to use the same sorts of
customizations I have been using with Emacs lo these many years.
If I need /two/ editors in order to:
a) wrap the Lisp environment, and
b) read news and mail,
I'll seriously consider dropping the one with the "marginally more
convenient Lisp environment" in favor of the one that does /all/ the
stuff I want it to do.
My set of applications are different from yours and from others'; the
result is that practically everyone will have reason to look at a
"Elisp-less CL-Emacs" as being functionally inferior to GNU Emacs /
XEmacs, leading to CL-Emacs getting ignored.
And the real point is that having two /nearly/ similar editors is much
/worse/ than having two very /different/ ones.
I think the Rice folks were /very/ right when they built the DrScheme
environment as having an editor totally different from Emacs. It's
distinct, so it can coexist without too much trouble.
> I think that if a cl-emacs "takes off" it will be on its own merits,
> not as a GNU Emacs clone. I'm not sure what the driving force will
> be; perhaps good integration with the development environment and
> CLIM, perhaps real editing with shared kill ring in every text
> widget.
The /huge/ win would be in performance, of having something that is:
a) Faster;
b) Truly multithreaded;
c) If it uses CLOS to structure control of the editor, that gives
some big wins in terms of typing and data validation...
>> I went through this with hemlock. Hemlock had Lisp interaction to
>> die for: at the time ilisp was a bad joke which would regularly
>> crash emacs, and even when it didn't hardly worked, but hemlock's
>> integration with CMUCL was a delight. I even wrote some clever
>> form-selection stuff in Hemlock, based on what TEdit did, which
>> would have been significantly more work in Emacs (then). But that
>> was *all* Hemlock did - it wouldn't run news, talk to CVS, keep
>> file headers up to date &c &c. And every 3 days some new cool
>> package would appear for Emacs, but not for Hemlock. Eventually I
>> stopped using it, when lemacs started working and supported X
>> properly, unlike emacs 18.
>
> This doesn't really explain why you stopped using Hemlock... was
> unbearable to switch between editors for different tasks?
I would think his second last sentence is a pretty good summary of
"why." /Nothing/ new comes out for Hemlock, but lots gets developed
for Emacs...
--
(concatenate 'string "cbbrowne" ·@cbbrowne.com")
http://www.ntlug.org/~cbbrowne/lsf.html
"Linux and other OSS advocates are making a progressively more
credible argument that OSS software is at least as robust -- if not
more -- than commercial alternatives." - Microsoft lamenting Open
Source Software in the "Halloween Document"
Christopher Browne <········@acm.org> writes:
> As far as /I/ am concerned, the vital things that /I/ want are the
> ability to run Gnus more efficiently, to edit SGML documents more
> efficiently, and to otherwise be able to use the same sorts of
> customizations I have been using with Emacs lo these many years.
>
> If I need /two/ editors in order to:
> a) wrap the Lisp environment, and
> b) read news and mail,
>
> I'll seriously consider dropping the one with the "marginally more
> convenient Lisp environment" in favor of the one that does /all/ the
> stuff I want it to do.
I'm still very torn about the whole Hemlock issue. I've been using
Emacs/ILISP recently for practical reasons (no CMUCL, thus no Hemlock
on Mac OS X, CLISP, OpenMCL...), but one really nice thing about
Hemlock is that it's integrated into the Lisp environment so much
better. This has a big win if you write your own special-purpose
development tools because you can integrate them easily with the
editor. On the other hand, I use GNU Emacs for *everything* else
(except Smalltalk).
> The /huge/ win would be in performance, of having something that is:
> a) Faster;
> b) Truly multithreaded;
On this end, I don't understand why the Emacs people don't make this a
priority. With the GNU code-generation library, and the reduced
number of weird, old platforms supported in Emacs 21, it should be
possible to make a bytecode JIT for most supported platforms, and make
a generational GC for all of them. Multiprocessing support would be
hard though. Not impossible (especially if you just went for
elisp-land stack groups), but hard.
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
"Thomas F. Burdick" <···@conquest.OCF.Berkeley.EDU> wrote in message
····················@conquest.OCF.Berkeley.EDU...
> I'm still very torn about the whole Hemlock issue. I've been using
> Emacs/ILISP recently for practical reasons (no CMUCL, thus no Hemlock
> on Mac OS X, CLISP, OpenMCL...), but one really nice thing about
> Hemlock is that it's integrated into the Lisp environment so much
> better. This has a big win if you write your own special-purpose
> development tools because you can integrate them easily with the
> editor. On the other hand, I use GNU Emacs for *everything* else
> (except Smalltalk).
So, I was looking at the editor in LispWorks, and at the source to Hemlock,
and at a 50,000 foot glance, they seem pretty similar. Anyone know if they
have similar roots?
Regards,
Will Hartung
(·····@msoft.com)
Christopher Browne <········@acm.org> writes:
> In an attempt to throw the authorities off his trail, Timothy Moore
<·····@pac-dhcp-22-84.amazon.com> transmitted:
> > Tim Bradshaw <···@cley.com> writes:
>
> As far as /I/ am concerned, the vital things that /I/ want are the
> ability to run Gnus more efficiently, to edit SGML documents more
> efficiently, and to otherwise be able to use the same sorts of
> customizations I have been using with Emacs lo these many years.
>
I don't think you'll ever get that from a CL-emacs, at least not one
that has GNU emacs compatibility as a specific goal.
> If I need /two/ editors in order to:
> a) wrap the Lisp environment, and
> b) read news and mail,
>
> I'll seriously consider dropping the one with the "marginally more
> convenient Lisp environment" in favor of the one that does /all/ the
> stuff I want it to do.
I really don't get this, but then I gave up reading mail in Emacs
some time ago.
>
> My set of applications are different from yours and from others'; the
> result is that practically everyone will have reason to look at a
> "Elisp-less CL-Emacs" as being functionally inferior to GNU Emacs /
> XEmacs, leading to CL-Emacs getting ignored.
>
> And the real point is that having two /nearly/ similar editors is much
> /worse/ than having two very /different/ ones.
If your explicit goal is to take over the world with an editor,
perhaps. But not if your goal is to write an editor that is
well-integerated with CL and CLIM, easily embedded in applications,
and extendable with CL.
>
> I think the Rice folks were /very/ right when they built the DrScheme
> environment as having an editor totally different from Emacs. It's
> distinct, so it can coexist without too much trouble.
>
> > I think that if a cl-emacs "takes off" it will be on its own merits,
> > not as a GNU Emacs clone. I'm not sure what the driving force will
> > be; perhaps good integration with the development environment and
> > CLIM, perhaps real editing with shared kill ring in every text
> > widget.
>
> The /huge/ win would be in performance, of having something that is:
> a) Faster;
> b) Truly multithreaded;
> c) If it uses CLOS to structure control of the editor, that gives
> some big wins in terms of typing and data validation...
The huge win would be not having to write Emacs lisp :)
Tim
Timothy Moore <·····@bricoworks.com> writes:
> Christopher Browne <········@acm.org> writes:
>
> >
> > The /huge/ win would be in performance, of having something that is:
> > a) Faster;
> > b) Truly multithreaded;
> > c) If it uses CLOS to structure control of the editor, that gives
> > some big wins in terms of typing and data validation...
>
> The huge win would be not having to write Emacs lisp :)
>
> Tim
But we are going to get that with giule.<shuder>
Just kidding, I hope.
marc
Christopher Browne <········@acm.org> writes:
> If I need /two/ editors in order to:
> a) wrap the Lisp environment, and
> b) read news and mail,
Hmm, when I used to use GNU Emacs for CL editing, I usually started
several separate emacs images anyway, so the problem is mainly one of
different interfaces, right?
The only two real problems I have with the Lispworks editor is:
- I hate that c-x 2 brings up a second window It's a bit distracting
- that I have to switch to my GNU Emacs window to do CVS interaction
on my lisp files.
I can live with these small problems, since the tight integration with
the development environment and its debugging tools is of really great
value to me.
--
(espen)
* Timothy Moore wrote:
> I'm not sure why any of this matters. Is Emacs really the only tool
> you use? Would you disregard a newreader out of hand if it didn't
> integrate with emacs?
If I ever intended to post articles, and if I had a choice, yes. If
I'm going to type at it, then it needs to support the tools to help me
that I'm used to, and in a way that I'm used to[1]. I don't have time
to waste learning some new, incompatible, text editing system.
> This doesn't really explain why you stopped using Hemlock... was
> unbearable to switch between editors for different tasks?
Yes, it was a complete saga. For instance, if I'm editing a file
under source control in emacs, and I want to check in a version, I
type C-xC-q and it just works - the header gets updated (partly by
emacs, partly by the source control system (emacs has to remember to
reread the file of course), I get prompted for a commit string, and so
on and so on. If I'm editing in hemlock, well, I can get a shell,
remember which source control system I'm using, type the right
command, get an editor (which won't be hemlock) to edit the commit
string, then go back to hemlock again, manually reread the file, and
I'm done. Except, bugger, it didn't do emacs's part of updating the
header. Oh well, I'll just live with a partly out of date header (or,
in practice, stop using file headers).
This is one small example of something I do *many* times a day, which
is maybe 20 times slower without the emacs support. Of course I could
*write* hemlock support for CVS/RCS/ClearCase (I don't need to use
SCCS fortunately, but I probably will need perforce sometime), but,
well, I have other things to do, things I get paid for for instance.
--tim
Footnotes:
[1] As well as, for instance, interacting with my mail reader in a
sane way - if I file an article in a mail folder I don't want it
to cause VM to have a seizure, say.
From: Hannah Schroeter
Subject: Re: CLEmacs [was: I'm a Lisper, hear me roar...]
Date:
Message-ID: <aqgvo8$70e$1@c3po.schlund.de>
Hello!
Erik Naggum <····@naggum.no> wrote:
>[...]
> right, Emacs is not just internals. What would /really/ make a difference
> to users would be if Emacs Lisp could be compiled to native code and not
> have to run through the byte-code interpreter.
Gnus would be really usable, for example. Speed is one major reason
why I'm using trn instead of Gnus.
>[...]
> but the amount of work to get there would simply be enormous.
That is right of course, and so I don't see any CL based Emacs or
Emacs like thing being in really widespread usage anytime soon.
Kind regards,
Hannah.
From: Michael Hudson
Subject: Re: I'm a Lisper, hear me roar... [was Re: Conference moment: Lisp certification?]
Date:
Message-ID: <7h3iszb27g7.fsf@pc150.maths.bris.ac.uk>
Kenny Tilton <·······@nyc.rr.com> writes:
> Michael Hudson wrote:
>
> > Kenny Tilton <·······@nyc.rr.com> writes:
> >
> > > I told my Franz rep y'all should send a thick-skinned, genial,
> > > proficient Lisper who is also Python-proficient into c.l.python to
> > > harrass them regularly with "why don't you use Lisp?".
> >
> > All that would do is fuck people off, I'd be willing to bet. People
> > don't like to be told they're doing things wrong, even if they are.
>
> OK, I should not have said "harrass", that was meant ironically, and I do
> not know of an emoticon for that.
>
> "Why don't you use Lisp?" should only be asked when it really solves the
> person's problem.
OK, I apologise for taking what you wrote at face value. Sarcasm
doesn't work if you have to label it as such, does it? Maybe that's
why there's no emoticon for it.
> Folks may still take exception to well-meaning advice, but that would be
> their problem, not the advice giver's. Any debate would either prove the
> Lisper wrong (telling Franz what to do next) or drag on long enough to
> persuade one Pythonian in a hundred to give Lisp a look.
This could well work -- I got interested in Haskell after mentions on
c.l.py (I was already interested in Lisp).
> Because you are right, most folks aren't looking to change languages
> at the drop of a remark. btw, from what I hear the python NGs are
> pretty cool, I am guessing that well-informed recommendations of
> Lisp would meet with a considered and considerate reception.
The Python NG was cooler before it got so busy[0], but I'd hope you're
right.
Cheers,
M.
[0] more busy -> less fun seems to be unavoidable, but it's a shame.
--
I have gathered a posie of other men's flowers, and nothing but
the thread that binds them is my own. -- Montaigne
From: Christopher C. Stacy
Subject: Re: Conference moment: Lisp certification?
Date:
Message-ID: <uy988a6ly.fsf@dtpq.com>
>>>>> On 04 Nov 2002 11:12:51 +0000, Tim Bradshaw ("Tim") writes:
Tim> You are right. However I think the reason CL gets laughed about is
Tim> not really lack of certification, it's a deeper malaise.
I would be surprised if people laugh about Lisp, because I
am sure they do not think about Lisp at all!
>>>>> On Mon, 04 Nov 2002 18:55:15 GMT, Kenny Tilton ("Kenny") writes:
Kenny> Yeah. I am jumping in here just to remind everyone that this came up
Kenny> at a conference during which the issue of Lisp's popularity came up
I wasn't there, but the impression I've gotten is that some fellow
who claimed to have some particular role in a Fortune 500 company
made a random remark about certification.
Given that major companies have used Lisp in the past, and are still
using it today, I am left wondering how seriously to take that remark,
especially as a key point in a larger context.
I am pretty sure that every Fortune 500 company uses Lisp in some
places in their IT organization. For example, they use Emacs.
We're trying to talk about Lisp's use as a principal technology,
but I bring up Emacs to illustrate that the question is one of context.
(Other "tool" context examples would be AutoCAD, or even Gensym.)
The main IT context of most Fortune 500 companies is absolutely not
based around technical aesthetics or the power that can come from there.
Large companies are usually trying to "solve" rather simple problems,
and the technology that they are interested in is one of plug-compatibility.
This means that the software, hardware, and personnel are all plug-compatible.
Just as a side note, there are major companies that use Lisp as a
core technology. In the late 1980s and early 1990s, the big success
stories for Lisp were all about Fortune 500 companies, especially
in the transportation and communications sectors. For example,
major phone companies used Lisp for fraud detection and other
applications. At one point, every American Express green card
transaction was approved by a Symbolics computer. The military has
also been a major user of Lisp, especially for scheduling.
Then there are all the proptietary AI applications,
some of which are still in use, and some passed.
At least one of the major airlines (but I think there were two)
uses Lisp for their flight planning (from strategic research, at the
real-time flight operations desks, all the way to the schedules the
passangers see on the monitors in the airport). I haven't checked
on this last example in three or four years, though, so it might
have been replaced. I know that the airline was about to do a major
upgrade of their system and was concerned that since Lisp was no
longer popular, they might not be able to find any programmers!
Notice that in the places where Lisp has been most successful, the
programmers have been experts and the problems have been "impossible".
Kenny> So "do we want F500 acceptance" was a given,
Kenny> we do not need to bat that around
The kind of acceptance from large companies that you are talking
about does not come as a result of the existance of some stupid
(or wonderful) certification test.
The acceptance criteria is whether every single school produces plenty
of graduates who actually program in Lisp. The company's requirement
is not that the programmer's be certified in a particular technology.
Rather, the company must be able to freely replace its programmers
with new hires who are knowledgable about the programming environment.
The other side of the requirement is that the company should not be
doing anything risky with their IT -- they need to be doing the same
standard thing as everybody else.
The purpose of the certification is to have some kind of filter,
even though it's very low grade, because there are hundreds of
thousands of programmers who claim to be replacement-able workers.
Most large companies are not doing anything interesting: they are
solving simple problems for which there are well-known solutions.
There isn't much creativity, and they are certainly not doing
research. What they are interested in is deploying the standard
solution, using the standard tools from the standard vendors,
and hiring the standard people.
Companies that actually have difficult problems (and realize that),
or ones that are doing research, are already willing to look at
"alternative" technologies, and they can generally be convinced to
try things like some of the newer functional languages, Smalltalk,
and even Lisp. But that's not the mainstream of industry.
Of course we would all like there to be lots of jobs where we can use
our skills, and we know that our programming language can solve even
pedestrian problems in a way that is often simpler and that scales
better for increasing complexity and evolutionary requirements.
We are frustrated at being forced to do projects in a slow and
painful way.
Establishing a filter, especially one that can be overcome by
memorizing some answers, isn't going to solve the problem that
there are not enough Lisp programmers to filter.
Making Lisp popular is a marketing problem, so thinking like a
technical person isn't going to get very far. You need to think
about the corporate requirements, which drive that market, and
then work back from there. (If you have a half-billion dollars
and some clout to spend on inflicting Lisp on the market, that's
another approach.)
The success of Java is about the success of having a zillion people
learning it in school, because it was shoved down people's throats
and hyped with a killer app (even though it wasn't much used for it),
and it was all about creating the conditions for the feedback that
brings up an environment "Learn To Be a Java Idiot In 2 Weeks".
Because that's what addresses the context of the corporate requirements.
······@dtpq.com (Christopher C. Stacy) writes:
> I am pretty sure that every Fortune 500 company uses Lisp in some
> places in their IT organization. For example, they use Emacs.
> We're trying to talk about Lisp's use as a principal technology,
> but I bring up Emacs to illustrate that the question is one of context.
> (Other "tool" context examples would be AutoCAD, or even Gensym.)
The primary consumers of Gensym's G2 are Fortune 50 companies.
So there.
:-)
* Christopher C Stacy wrote:
> At least one of the major airlines (but I think there were two)
> uses Lisp for their flight planning (from strategic research, at the
> real-time flight operations desks, all the way to the schedules the
> passangers see on the monitors in the airport). I haven't checked
> on this last example in three or four years, though, so it might
> have been replaced. I know that the airline was about to do a major
> upgrade of their system and was concerned that since Lisp was no
> longer popular, they might not be able to find any programmers!
One of the airlines that did was swissair, I think. Oh, look what
happened to them!
(*not* because of Lisp of course, but I thought it was amusing)
--tim
> I wasn't there, but the impression I've gotten is that some fellow
> who claimed to have some particular role in a Fortune 500 company
> made a random remark about certification.
All correct except `random remark' --- he seemed to feel that Lisp
simply would not succeed in the corporate world without
certification.
--
Fred Gilham ······@csl.sri.com
America does not know the difference between sex and money. It treats
sex like money because it treats sex as a medium of exchange, and it
treats money like sex because it expects its money to get pregnant and
reproduce. --- Peter Kreeft
* Christopher C Stacy wrote:
> The acceptance criteria is whether every single school produces plenty
> of graduates who actually program in Lisp. The company's requirement
> is not that the programmer's be certified in a particular technology.
> Rather, the company must be able to freely replace its programmers
> with new hires who are knowledgable about the programming environment.
> The other side of the requirement is that the company should not be
> doing anything risky with their IT -- they need to be doing the same
> standard thing as everybody else.
> [...]
> Establishing a filter, especially one that can be overcome by
> memorizing some answers, isn't going to solve the problem that
> there are not enough Lisp programmers to filter.
However, we need something to ensure that the people the schools (this
is USAan for `universities' isn't it?) produce who say they can write
Lisp actually *can* write Lisp. Right now, a lot of people who claim
to know Lisp come from academic AI backgrounds where they have been
taught what little they know by my `knew a little maclisp once'
academic, and combine the kind of breathtaking arrogance which is
unfortunately common in AI with the ability to write completely
incomprehensible and inefficient code. The results of these kind of
people working in industry are a lot of bad feelings about Lisp.
I'm not making this up, by the way. I spent nearly a year from late
2000 working on a deployed system in a telco where one or more of
these people had worked. It took me most of that year to convince
people that Lisp didn't have to be this bad and that not all Lisp
people were arrogant shits.
I don't think a filter, in the certification sense, would help much
(though I can think of some nice questions which would detect many of
the more bogus people!). I *do* think a more engineering-based and
less AI-academic approach to teaching Lisp would help a lot.
--tim
Tim Bradshaw wrote:
>
>
> I don't think a filter, in the certification sense, would help much
> (though I can think of some nice questions which would detect many of
> the more bogus people!).
Care to share? But good exam questions are hard to craft, so understood if
you pass.
* Kenny Tilton wrote:
> Care to share? But good exam questions are hard to craft, so
> understood if you pass.
you're right about crafting things. But questions which encouraged
easy-in-Lisp but algorithmically terrible solutions; `trick' questions
which encouraged people to smash the CL package; any question about
the package system or the error system; others I'm sure.
--tim
On Tue, 5 Nov 2002, Christopher C. Stacy wrote:
> The success of Java is about the success of having a zillion people
> learning it in school, because it was shoved down people's throats
> and hyped with a killer app (even though it wasn't much used for it),
> and it was all about creating the conditions for the feedback that
> brings up an environment "Learn To Be a Java Idiot In 2 Weeks".
> Because that's what addresses the context of the corporate requirements.
I agree, but there's one thing missing from your list of superficial
attributes. Namely, the "net" hype that went along with Java. Without the
Java hype, we still would have had the "net" hype. But without the "net"
hype, the Java hype would have been another blip on the programming
languages radar screen. Maybe a 747 rather than a sparrow (though I doubt
it), but a blip nevertheless. And of course it wasn't ordained to work out
that way; Sun was lucky and successfuly opportunistic.
* Nick Geovanis
| IT Computing Svcs Computing's central challenge:
| Northwestern Univ How not to make a mess of it.
| ··········@nwu.edu -- Edsger Dijkstra
+------------------->
From: Christopher C. Stacy
Subject: Re: Conference moment: Lisp certification?
Date:
Message-ID: <uof93z24p.fsf@dtpq.com>
>>>>> On Tue, 5 Nov 2002 16:06:17 -0600, Nicholas Geovanis ("Nicholas") writes:
Nicholas> On Tue, 5 Nov 2002, Christopher C. Stacy wrote:
>> The success of Java is about the success of having a zillion people
>> learning it in school, because it was shoved down people's throats
>> and hyped with a killer app (even though it wasn't much used for it),
>> and it was all about creating the conditions for the feedback that
>> brings up an environment "Learn To Be a Java Idiot In 2 Weeks".
>> Because that's what addresses the context of the corporate requirements.
Nicholas> I agree, but there's one thing missing from your list of superficial
Nicholas> attributes. Namely, the "net" hype that went along with Java. Without the
Nicholas> Java hype, we still would have had the "net" hype. But without the "net"
Nicholas> hype, the Java hype would have been another blip on the programming
Nicholas> languages radar screen. Maybe a 747 rather than a sparrow (though I doubt
Nicholas> it), but a blip nevertheless. And of course it wasn't ordained to work out
Nicholas> that way; Sun was lucky and successfuly opportunistic.
The "net" is the killer app to which I was referring.
On Wed, 6 Nov 2002, Christopher C. Stacy wrote:
> >> On Tue, 5 Nov 2002, Christopher C. Stacy wrote:
> >> The success of Java is about the success of having a zillion people
> >> learning it in school, because it was shoved down people's throats
> >> and hyped with a killer app (even though it wasn't much used for it),
> >> and it was all about creating the conditions for the feedback that
> >> brings up an environment "Learn To Be a Java Idiot In 2 Weeks".
> >> Because that's what addresses the context of the corporate requirements.
>
> Nicholas> I agree, but there's one thing missing from your list of superficial
> Nicholas> attributes. Namely, the "net" hype that went along with Java. Without the
> Nicholas> Java hype, we still would have had the "net" hype. But without the "net"
> Nicholas> hype, the Java hype would have been another blip on the programming
> Nicholas> languages radar screen. Maybe a 747 rather than a sparrow (though I doubt
> Nicholas> it), but a blip nevertheless. And of course it wasn't ordained to work out
> Nicholas> that way; Sun was lucky and successfuly opportunistic.
>
> The "net" is the killer app to which I was referring.
But not all megacorps meant the same thing when saying "net".
Sun marketing did their level best to equate Java and "net" in
people's minds. On the other side AOL, M$oft, IBM, Oracle and 10000
different ISPs were either hedging their bets (like IBM and Oracle),
countering Java as best they could (M$oft), or simply pushing their own
view of "net" (ISPs) which neither confirmed nor denied Sun's.
But I haven't heard much about IBM's "7500 Java programmers" lately,
how about you?
* Nick Geovanis
| IT Computing Svcs Computing's central challenge:
| Northwestern Univ How not to make a mess of it.
| ··········@nwu.edu -- Edsger Dijkstra
+------------------->
Will Hartung wrote:
> If you think about it, the only real leap that can be made from Java is into
> something like Lisp, or maybe a functional language (but I'm not familiar
> enough with those to say).
A more obvious leap would be Java->Smalltalk.
> Will Hartung
Le Hibou
--
Dalinian: Lisp. Java. Which one sounds sexier?
RevAaron: Definitely Lisp. Lisp conjures up images of hippy coders,
drugs,
sex, and rock & roll. Late nights at Berkeley, coding in Lisp fueled by
LSD.
Java evokes a vision of a stereotypical nerd, with no life or social
skills.
Kenny Tilton <·······@nyc.rr.com> writes:
[...]
> -- unwind-protect
>
> -- exceptions))
>
> OK, that looks like a TOC for any of the Lisp books out there,
[...]
Could you point me at some decent Lisp books that cover Lisp error
(sorry, condition) handling ? In my personal library the only one at
the moment is CLtL2 (... and KMPs' paper:
http://world.std.com/~pitman/Papers/Condition-Handling-2001.html).
I'd be interested in something that is a bit more turorial-like, a bit
less abstract than the coverage in the above. Admittedly, my problem
is largely, finding enough uniterrupted time to concentrate on the
material, but I can't help feeling that a different approach to the
subject could help me overcome the initial barrier more easily.
Any suggestions ?
Thanks,
Jacek
Jacek Generowicz <················@cern.ch> writes:
> Kenny Tilton <·······@nyc.rr.com> writes:
>
> [...]
>
> > -- unwind-protect
> >
> > -- exceptions))
> >
> > OK, that looks like a TOC for any of the Lisp books out there,
>
> [...]
>
> Could you point me at some decent Lisp books that cover Lisp error
> (sorry, condition) handling ?
AFAIK, there aren't any. However ...
> In my personal library the only one at the moment is CLtL2 (... and
> KMPs' paper:
> http://world.std.com/~pitman/Papers/Condition-Handling-2001.html).
... Kent said that he was working on (or thinking of working on) a
Lisp book. I'd imagine he'd cover the condition system, so at some
point ther might be one :)
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
Jacek Generowicz <················@cern.ch> writes:
> Could you point me at some decent Lisp books that cover Lisp error
> (sorry, condition) handling ? In my personal library the only one at
> the moment is CLtL2 (... and KMPs' paper:
> http://world.std.com/~pitman/Papers/Condition-Handling-2001.html).
This might also be good:
http://world.std.com/~pitman/Papers/Exceptional-Situations-1990.html
--
Janis Dzerins
If million people say a stupid thing, it's still a stupid thing.
Jacek Generowicz <················@cern.ch> writes:
>
> Could you point me at some decent Lisp books that cover Lisp error
> (sorry, condition) handling ?
To any would-be authors reading this, a tutorial on real life
of the condition system (especially restarts!) would be a
great addition to the online CL cookbook.
Kenny Tilton <·······@nyc.rr.com> writes:
;; > already), then big companies use it, and our choices are now bigger.
;;
;; Yeah. I am jumping in here just to remind everyone that this came up
;; at a conference during which the issue of Lisp's popularity came up
;; again and again. And again. Then there was a fishbowl on it. Then we
;; broke for snacks and talked about it.
;;
;; So "do we want F500 acceptance" was a given, we do not need to bat
;; that around. This is not to say that some of us would rather starve
;; than work for f500 (or get certified or take a lie-detector test
;; or
I wouldn't spend a farts worth of time or effort for Microsoft
certification. It's a lousy product and the certification is bound
to involve too much point-an-click time. I'd spend a good deal of
time and effort getting certified in CL. Not all certifications are
equal and, I think, the worst case scenarios are tending to be voiced
here.
;; give a urine sample or whatever), just that enough Lispers want f500
;; acceptance that we can now look at a second question: how much would
;; certification matter? We have one data point from Mr. megaCorp. Any
;; others?
In my experience the biggest hurdle to change in any organizational
infrastructure isn't assesment and analysis of technologies, but
adoption and implementation costs: it's not whether or not Lisp works
better than other languages, but what's involved in getting Lisp
onsite. Besides the hard and fast costs associated with such a move
there may be, in this case, ambiguity about standards and efficacy.
Often, the clearly defined, but sub-optimal, solution is favored over
the clearly optimal end-point with an unclear path. I think
certification would go a long way to ammeliorating that ambiguity.
;; At the same time, i suddenly find myself intrigued by the question of
;; what would we want on a Lisp certfication exam?
Why does it have to be an 'exam'? Exams (of the type I'm thinking of,
and which may or may not map to your idea) are pretty horrible ways
of assessing skillsets and are really only useful as reducing the
workload of the test giver. Lispers are, by and large, a pretty
creative and fun-loving bunch... why not apply that to the solution
space of certification??
Eriks' idea of a paper/presentation is close to an oral exam which is
one of the better ways of assessing skillsets *and* ability to
articulate difficult and/or subtle concepts. The burden, however,
placed on the test-giver is scaled up many orders of magnitude.
My own suggestion would take the outward form of a contest, but
throwing out the 'winner-take-all' mode[1] in favor of a point system
and defining a strict set of guidelines for tackling a given problem.
Pose a problem then state the problem-solving space:
An example problem: write a program to translate the age
of your dog back and forth between human years and dog
years using Roman Numerals, where:
(example problem-solving space)
You must use CL
You must demonstrate knowledge of
CLHS
ANSI standards
CLOS
MACROS
... etc ...
You must comment your code properly
... etc ...
Points will be deducted for bad Urdu and added for good Sanskrit. The
certifiers (judges) will 'grade' and your certification results from
your grade.
Peace,
Petr
[1] A side effect would be to parrallelize the cert with an
actual winner-take-all contest to attract attention and competition.
* Petr Swedock wrote:
> Why does it have to be an 'exam'? Exams (of the type I'm thinking of,
> and which may or may not map to your idea) are pretty horrible ways
> of assessing skillsets and are really only useful as reducing the
> workload of the test giver. Lispers are, by and large, a pretty
> creative and fun-loving bunch... why not apply that to the solution
> space of certification??
Reducing the workload of the test-giver matters, if their time costs
money. And if Lisp is an in-demand skill, their time will cost quite
a lot of money. In general you need to make sure that whatever test
you want to have (if you want one) is affordable, because money won't
magically appear to fund it. For certification-type tests, generally
the test-taker pays for the test, and if it's a lot of money, not many
people will take he test...
(This goes for anything, of course: for instance we can't just say
`let's standardise x bit of functionality', we need to find the time -
and therefore money - to actually get it done.)
--tim
On Mon, 04 Nov 2002 18:55:15 GMT, Kenny Tilton <·······@nyc.rr.com> wrote:
> Yeah. I am jumping in here just to remind everyone that this came up at
> a conference during which the issue of Lisp's popularity came up again
> and again. And again. Then there was a fishbowl on it. Then we broke for
Wasn't the--international--conference itself a strong hint that Lisp has
enough popularity? By the way, how many people attended ILC 2002?
Paolo
--
EncyCMUCLopedia * Extensive collection of CMU Common Lisp documentation
http://www.paoloamoroso.it/ency/README
Paolo Amoroso wrote:
> On Mon, 04 Nov 2002 18:55:15 GMT, Kenny Tilton <·······@nyc.rr.com> wrote:
>
>
>>Yeah. I am jumping in here just to remind everyone that this came up at
>>a conference during which the issue of Lisp's popularity came up again
>>and again. And again. Then there was a fishbowl on it. Then we broke for
>
>
> Wasn't the--international--conference itself a strong hint that Lisp has
> enough popularity? By the way, how many people attended ILC 2002?
>
I myself was encouraged by the size of the '02 conference, esp. when
compared with the one in '99.
But this may have been due to the extraordinary efforts of Ray and the
ALU and Franz. And because they cast the net wider; I heard a lot of
talks that included "I suppose I should mention Lisp at least once
before ending, so...we use Lisp. Questions?".
:)
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
>>>>> "Paolo" == Paolo Amoroso <·······@mclink.it> writes:
Paolo> By the way, how many people attended ILC 2002?
I have heard the number ~120 including the 40-50 speakers.
It was my first lisp conference so I do not have anything to compare
with, but being there certainly gave me a warm fuzzy feeling,
testifying to the continued viability of LISP.
------------------------+-----------------------------------------------------
Christian Lynbech | Ericsson Telebit, Skanderborgvej 232, DK-8260 Viby J
Phone: +45 8938 5244 | email: ·················@ted.ericsson.se
Fax: +45 8938 5101 | web: www.ericsson.com
------------------------+-----------------------------------------------------
Hit the philistines three times over the head with the Elisp reference manual.
- ·······@hal.com (Michael A. Petonic)
"J L Russell" <·········@alum.mit.edu> wrote in
····························@bgtnsc04-news.ops.worldnet.att.net:
>
> "Tim Bradshaw" <···@cley.com> wrote in message
> ····················@cley.com...
>> * J L Russell wrote:
>> > And all I'm saying is, do we really want to work for the forces of
>> > darkness?
>>
>> We'd like to get paid, yes.
>>
>> --tim
>
> So, Philip Morris, RJR, BAT, etc. approaches you (pl.) and says,
> "Look, we know that lisp is the most effective tool around.
> We want you to design a system that consolidates all our
> research on increasing addictiveness of cigarettes and
> marketing them to children in order to help us design
> a more effective strategy."
> You think, wow, we're so desperate, we'll take anything.
>
> I don't know about you (pl.), but I'd rather go on the dole.
> My utility to society would be _increased_ that way.
>
> -James Russell
>
>
>
Many people I know would call working fo a goverment the "forces of
darkness", for example working for the NSA, or Exxon. You had to find
the worst possable interpitation( it is illegal) of this and then use it
as a public insult. What inspired you to behave as such an ass?
marc
"Marc Spitzer" <········@optonline.net> wrote in message ·········································@167.206.3.3...
> "J L Russell" <·········@alum.mit.edu> wrote in
> ····························@bgtnsc04-news.ops.worldnet.att.net:
>
> >
> > "Tim Bradshaw" <···@cley.com> wrote in message
> > ····················@cley.com...
> >> * J L Russell wrote:
> >> > And all I'm saying is, do we really want to work for the forces of
> >> > darkness?
> >>
> >> We'd like to get paid, yes.
> >>
> >> --tim
> >
> > So, Philip Morris, RJR, BAT, etc. approaches you (pl.) and says,
> > "Look, we know that lisp is the most effective tool around.
> > We want you to design a system that consolidates all our
> > research on increasing addictiveness of cigarettes and
> > marketing them to children in order to help us design
> > a more effective strategy."
> > You think, wow, we're so desperate, we'll take anything.
> >
> > I don't know about you (pl.), but I'd rather go on the dole.
> > My utility to society would be _increased_ that way.
> >
> > -James Russell
> >
> >
> >
>
> Many people I know would call working fo a goverment the "forces of
> darkness", for example working for the NSA, or Exxon. You had to find
> the worst possable interpitation( it is illegal) of this and then use it
> as a public insult. What inspired you to behave as such an ass?
>
> marc
I'm not sure what your point is about the government and Exxon.
Change 'children' to 'adult' if you like, no longer illegal, not much less immoral.
Who did I insult?
Tim answered my trite, non-specific rhetorical question with what I saw as a non-serious,
non-specific answer.
I responded with another example, not aimed at anyone in particular,
which Tim admirably countered.
If you took it personally, that's not my problem.
If you happen to market cigarettes to children, I stand by what I said.
I apparently underestimated how seriously people take this.
I thought we were debating philosophy here, not flinging insults.
You were the first to call anyone in particular a name.
-James
"J L Russell" <·········@alum.mit.edu> wrote in
····························@bgtnsc04-news.ops.worldnet.att.net:
>
> "Marc Spitzer" <········@optonline.net> wrote in message
> ·········································@167.206.3.3...
>> "J L Russell" <·········@alum.mit.edu> wrote in
>> ····························@bgtnsc04-news.ops.worldnet.att.net:
>>
>> >
>> > "Tim Bradshaw" <···@cley.com> wrote in message
>> > ····················@cley.com...
>> >> * J L Russell wrote:
>> >> > And all I'm saying is, do we really want to work for the forces
>> >> > of darkness?
>> >>
>> >> We'd like to get paid, yes.
>> >>
>> >> --tim
>> >
>> > So, Philip Morris, RJR, BAT, etc. approaches you (pl.) and says,
>> > "Look, we know that lisp is the most effective tool around.
>> > We want you to design a system that consolidates all our
>> > research on increasing addictiveness of cigarettes and
>> > marketing them to children in order to help us design
>> > a more effective strategy."
>> > You think, wow, we're so desperate, we'll take anything.
>> >
>> > I don't know about you (pl.), but I'd rather go on the dole.
>> > My utility to society would be _increased_ that way.
>> >
>> > -James Russell
>> >
>> >
>> >
>>
>> Many people I know would call working fo a goverment the "forces of
>> darkness", for example working for the NSA, or Exxon. You had to
>> find the worst possable interpitation( it is illegal) of this and
>> then use it as a public insult. What inspired you to behave as such
>> an ass?
>>
>> marc
>
> I'm not sure what your point is about the government and Exxon.
> Change 'children' to 'adult' if you like, no longer illegal, not much
> less immoral. Who did I insult?
Tim specificly and everyone else by implacation. The "I am too good for
this but I do not know about the rest of you" in insulting.
> Tim answered my trite, non-specific rhetorical question with what I
> saw as a non-serious, non-specific answer.
> I responded with another example, not aimed at anyone in particular,
> which Tim admirably countered.
> If you took it personally, that's not my problem.
> If you happen to market cigarettes to children, I stand by what I
> said.
Now you are saying I market death to children, well *FUCK YOU*. If you
are too much of a childish dolt to be able to figure out that I was
commenting on your actions, by reading the text I wrote you twit, go take
a remedial reading class at the local jr college.
Since you stand by the crap you said you must be an asshole.
>
> I apparently underestimated how seriously people take this.
> I thought we were debating philosophy here, not flinging insults.
> You were the first to call anyone in particular a name.
You set the tone with your "market death to kids" example you waste of
space. Then you follow it up with an accusation that I market death to
kids. My father died when I was 11 from his 3rd heart attack and he
smoked. He was sick for almost all the time I remember him. I do not
need the dole not to work for a tobacco company.
marc
>
> -James
>
>
>
>
>
>
>
>
"J L Russell" <·········@alum.mit.edu> writes:
;;
;; And all I'm saying is, do we really want to work for the
;; forces of darkness?
;;
;; Any company that has the expertise to determine their true
;; needs and assess programmer competence doesn't need these tests.
;; Use of these tests is an admission of managerial incompetence.
;; Who wants to work for such a company?
That is not at all the case. There is incompetent management and there
is competent management that knows nothing about Lisp. Furthermore,
management that is competent would like to go about it's job without
having to learn anything about Lisp. That's what would make them
competent (or do you think Jack Welch knows anything about building
jet engines or MRI equipment????)
Certification is merely a means for one authority to trust another.
The OP mentions anecdotal evidence that adoption of Lisp as a problem
solving tool is hindered by lack of such certification... and hints
that lack of such tool is perceived as non-existance of authority.
What is your response to that??
;; I might do so only out of sheer desperation.
You may whine about how non-Lispers don't 'get it'. and how management
misses out on such rapturous experience as Lisp coding... but from
where they sit it might just look like an un-disciplined free-for-all
lacking adult supervision.
;; Is it possible that companies use these tests because they just
;; happen to measure exactly the skills that the company needs?
;; No. I have taken many tests in my life, and done quite well
;; on most all of them, and I have never seen one that
;; measures much of anything other than the ability to do well
;; on that particular test. I know that this is a trite
;; observance, but that doesn't make it any less true.
It's not trite. It's naive. And certification doesn't map directly
to a test.
Peace,
Petr
* Petr Swedock wrote:
> Certification is merely a means for one authority to trust another.
This is a good point. A certificate doesn't have to be some kind of
trivial thing. Many of us have a certificate that says something like
`this person went to such and such a university and was there for so
and so years and did these courses and passed some exams at the end of
it all, and we think they did OK' - a degree in other words. These
things are pretty widely used when deciding whether to employ people.
--tim
Tim Bradshaw <···@cley.com> wrote:
+---------------
| * Petr Swedock wrote:
| > Certification is merely a means for one authority to trust another.
|
| This is a good point. A certificate doesn't have to be some kind of
| trivial thing. Many of us have a certificate that says something like
| `this person went to such and such a university and was there for so
| and so years and did these courses and passed some exams at the end of
| it all, and we think they did OK' - a degree in other words. These
| things are pretty widely used when deciding whether to employ people.
+---------------
Hmmm... You guys may be onto something here. What if instead of a
"certification exam" per se one simply reported one's grades in one
of several respected ("certified") extension courses in "industrial
strength" Lisp programming? Then the burden of "certification" falls
on the schools teaching those courses (as it does with universities).
[Of course, you still have to interest such schools in *teaching* such
courses, and in hiring decent-enough teachers that a passing grade has
some meaning...]
-Rob
-----
Rob Warnock, PP-ASEL-IA <····@rpw3.org>
627 26th Avenue <URL:http://www.rpw3.org/>
San Mateo, CA 94403 (650)572-2607
* Rob Warnock wrote:
> [Of course, you still have to interest such schools in *teaching* such
> courses, and in hiring decent-enough teachers that a passing grade has
> some meaning...]
That's something I'd encourage - Lisp should be taught by the people
who currently teach Ada, not by those who teach AI...
--t
From: Christopher C. Stacy
Subject: Re: Conference moment: Lisp certification?
Date:
Message-ID: <uvg3e8i5h.fsf@dtpq.com>
>>>>> On Sun, 03 Nov 2002 16:56:17 GMT, Kenny Tilton ("Kenny") writes:
Kenny> As it stands, CL gets laughed out of the running in IT, because the
Kenny> immune system identifies us as "not-ready-for-IT".
Thing is, I am not at all sure that I believe this claim.
Kenny Tilton <·······@nyc.rr.com> writes:
> I have, and I can't say I heard anyone talking about
> certification. otoh, i did know one IT recruiter who thought some guy
> was a genius because he had passed a wadge of MS exams in short order
> (an Asian!).
I don't quite understand this repeated reference to "Asians" with
regards to certification. What is the significance? I have worked
with "Asians" of various backgrounds and nationalities, both within
the U.S. and abroad, and I'm unable to discern this connection that
keeps popping up here.
--
Sincerely,
Craig Brozefsky <·····@red-bean.com>
Free Scheme/Lisp Software http://www.red-bean.com/~craig
Craig Brozefsky wrote:
> Kenny Tilton <·······@nyc.rr.com> writes:
>
>
>>I have, and I can't say I heard anyone talking about
>>certification. otoh, i did know one IT recruiter who thought some guy
>>was a genius because he had passed a wadge of MS exams in short order
>>(an Asian!).
>
>
> I don't quite understand this repeated reference to "Asians" with
> regards to certification. What is the significance? I have worked
> with "Asians" of various backgrounds and nationalities, both within
> the U.S. and abroad, and I'm unable to discern this connection that
> keeps popping up here.
>
It's a deliberately politically incorrect joke derived from a laughing
assertion by -- you guessed it -- one of my best friends, an Asian, viz,
"We just memorize the book!"
Did you hear the one about the Ebonics spelling contest?
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
* Kenny Tilton wrote:
> As it stands, CL gets laughed out of the running in IT, because the
> immune system identifies us as "not-ready-for-IT". Maybe if we adorn
> ouselves with a few ITish proteins like certification the immune
> system will think we are friend, not foe.
You are right. However I think the reason CL gets laughed about is
not really lack of certification, it's a deeper malaise. The problem
is that there is a huge gap between the way Lisp is typically
presented and what CL is actually like:
CL: advanced but industrial language, designed by people with
experience of large real-world systems (like operating systems
deployed outside academia &c). Has rejected many `research
language' ideas in favour of clear implementability (hence: no
call/cc, no mandatory TRO, no standard MOP), and so on. Has left
in plenty of `hacky old' stuff so that existing programs could be
got to run. Has GO TO. Language purity a non-goal. CL is an
enormous compromise with the aim being an implementable system
which will support large programs. And it does *really well* at
this: better than any other language I know. CL is the Java of
Lisp. Programming language purists almost universally hate CL.
Lisp presentation: random AI academic who once knew a little bit
of maclisp waves their arms about Lisp and says a bunch of random
stuff, about 80% of which is wrong, and the other 40% is bogus AI
crap[1]. In the advanced course, students are shown a large
program written by this academic. The only data types it uses are
lists and symbols (except he (it is always a he) explains that
they are lists and atoms). It implements it's own rational number
package because sponge lisp didn't have rationals in 1979. It's
all in one package (if the package is mentioned it is called USER
not CL-USER). Error handling is done by writing the form to be
evaluated to a file, cranking up the system to load the file and
write output to another file, and checking this second file for
the string `Err'. The program is loaded by catting all the source
files together and loading that file. It can not be run compiled
(lisp has no compiler, after all), and makes prolific use of EVAL.
It runs on an old version of kcl, breaking on akcl or any more
recent Lisp. The indentation style is unconventional but is
presented as the correct way of indenting programs. Comments are
in Latin and bad Urdu.
After presenting this system, the academic goes on to present his
new system which has been written in C++ by some spare PhD
students he had lying around. Unfortunately although this system
is far better, it isn't quite finished yet and it tends to core
dump a lot (like: core dump is all it will actually do). Some
small soon-to-be-solved issues to do with memory allocation mean
that it currently uses rather more memory than the old system.
It's currently being reimplemented in Java, and the parts of it
which rely on cranking up the old system every time it needs to
actually do anything are being redone in perl. Comments are now
in middle english.
After the presentation of the new system, students who hang around
will be accosted by a particularly haggard looking PhD student,
who explains that the demo they actually saw was done by a
reimplementation of the system in a few hundred lines of modern CL
that a couple of the PhD students cooked up over a weekend. Over
half of this new, working, system consists of code to replicate
the idiosyncratic messages from the old system, in particular the
incorrect printing of numbers in English (they tried using ~R, but
it got things right). One of the students who implemented this new
system tried to present it to the academic, but emerged white &
shaking from his office, and left the university shortly
afterwards without completing a PhD. No one else has dared raise
the issue, so instead they struggle on with the Java/Perl version,
which is now 35,000 lines, and in the process of being converted
(again) to work in the current Java version.
You think I'm joking, don't you? I wish I was. If I'm a random
manager who might consider using Lisp, I'm likely either to remember
something like the above, or ask some minion who will. Because, like
it or not, this *is* how Lisp is presented in academia. Is it
surprising that it doesn't get used as much as it might do? We need
to do something about this.
I don't think certification is the answer, and neither do I think
giving papers (the academic has *lots* of published papers on his
system!) helps much. We need to change how CL is taught, and if we
can't do that (and I can't see a way other than killing most of the
current AI academics) we need to do a huge amount of work to present
CL as a serious, industrial system, so that it is easy to point people
at examples of Lisp being used industrially, and so they can find
books and web pages describing how well it supports this. *Then*
certification might help.
--tim
Footnotes:
[1] You think this doesn't add up? Read on...
Tim Bradshaw wrote:
> I don't think certification is the answer, and neither do I think
> giving papers (the academic has *lots* of published papers on his
> system!) helps much. We need to change how CL is taught, and if we
> can't do that (and I can't see a way other than killing most of the
> current AI academics) we need to do a huge amount of work to present
> CL as a serious, industrial system, so that it is easy to point people
> at examples of Lisp being used industrially, and so they can find
> books and web pages describing how well it supports this. *Then*
> certification might help.
I don't if this could be relevant but you may check out
http://www.cs.unca.edu/~manns/intropatterns.html for some food for thought.
Pascal
--
Pascal Costanza University of Bonn
···············@web.de Institute of Computer Science III
http://www.pascalcostanza.de R�merstr. 164, D-53117 Bonn (Germany)
* Tim Bradshaw <···@cley.com>
| I think the problem with this is that the skills needed to write
| conference papers are different than the skills needed to write
| significant programs.
The problem with all certifications is that the skills needed to get the
certification are different from the skills needed in the job requiring
the certification. The question we should ask "does this correlate with
what we actually need?" and allow ourselves to be surprised by the many
unexpected things that do correlate. The human brain is deficient in its
lack of capacity to see how many small things work together. We are very
good at singling out things that are important and focus on that one
thing, but lousy at keeping track of masses of smaller interests that
work together to make change. Usenet is an interesting experiment in
this regard. Some people are so unable to process more than one quality
at a time that they have to /invent/ aspects in order to retro-support
their favorite quality. This is not just the massive stupidity and lack
of intelligence it looks like, it is how people are naturally wired to
deal with the world if they do not consciously override it by thinking.
This is somewhat like voting for people to lead a country. The United
States is /really/ paying the price for its plurality system during this
presidential period. For some reason, how many people would like
something the most is regarded as a reason to choose it. I favor a
system where the number of people who like something the /least/ is
subtracted from the number of people who like it the most, or generally,
a system where each candidate is given positive and negative scores in
some small range (where the sum of the absolute value of all scores is
constant or has a fixed upper limit) and those you feel nothing about
gets zero or no vote at all. The scores are simply summed and whoever
gets the highest total score wins. The purpose of the negative votes is
to ensure that someone who may well be favored by the largest minority
but is loathed by a larger group, perhaps even a majority, not get into a
position where the majority would feel they were not heard and which
would destabilize the entire system. This would ensure that a candidate
would want to get backers on issues, not just fans of their person (or to
avenge their father), and would have to calculate the risk of offending
some groups, not just run over them.
Back to certification, the number of points at which you would have to
score well to be a good candidate for a job is attempted destilled into a
certification, which at best may be assumed to mean "above the baseline",
but the result may well be as undesirable as making George W. Bush the
Republican presidential candidate. In general, I want examinations and
tests to score negative for a wrong answer and zero for no answer, and if
it were up to me, "I don't know" would be far more socially acceptable
than "I guess". But no such luck. Even the business community favors
people who make wrong decisions over those who try to figure out what the
best thing is and effectively make a decision not to act. But you get
what you deserve when you operate that way. Unfortunately, people who
should not get a job in programming get one because of certification,
people who should be kept as far away from money as possible run both
Enron and WorldCom and Arthur Andersen into the ground, and people who
should be kept as far as away from Washington D.C. as possible get into
the White House instead of staying in Texas and lots and lots of people
suffer worldwide. Incompetence should be a criminal offence. The core
problem is that certification does not solve any big problems, only small
ones, just as book-keeping and auditing does not keep people from being
criminals.
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
+ Erik Naggum <····@naggum.no>:
| This is somewhat like voting for people to lead a country. The
| United States is /really/ paying the price for its plurality
| system during this presidential period. For some reason, how many
| people would like something the most is regarded as a reason to
| choose it. I favor a system where the number of people who like
| something the /least/ is subtracted from the number of people who
| like it the most, or generally, a system where each candidate is
| given positive and negative scores in some small range (where the
| sum of the absolute value of all scores is constant or has a fixed
| upper limit) and those you feel nothing about gets zero or no vote
| at all. The scores are simply summed and whoever gets the highest
| total score wins.
This is probably equivalent to a system in which every voter has a
fixed number of (positive) points to distribute among candidates as he
wishes: Just add a constant to the points awarded by each voter in
your system to see this. Such systems are prone the same kind of
paradoxes that plague all the more conventional systems in
existence. These paradoxes are similar to, but not the same as,
Arrow's theorem, which is often interpreted as saying that perfect
democracy is impossible.
Anyway, I think a better way to achieve what you want (though not
paradox free - no fair voting system can be paradox free) is the
single transferrable vote. Each voter ranks all candidates. If one
candidates is ranked first on more than half of the ballots, he
wins. Otherwise, the candidate who is ranked first on the smallest
number of ballots is thrown out of the race, all the ballots are
adjusted accordingly, and the procedure starts from the top. (If N
candidates are to be elected, the requirement for being elected is to
have at least a proportion 1/(N+1) of the votes. Also, the ballots
that helped elect a winning candidate are put back in the pile of
ballots, albeit with a reduced weight. This case gets a bit more
complex.) If the latest US presidential election had been run this
way, one can imagine many voters ranking the candidates Nader, Gore,
Bush. Presumably there would be enough Nader+Gore votes to stop Bush
from winning outright; then since Nader would probably have the
smallest set of votes, he would be out of the race, most of his votes
would have gone for Gore instead, and Gore would win.
With an election system like this, it would be impossible for the
candidate of the largest minority to win, if he is uniformly loathed
by all the other minorities.
To get this if not back on topic, or at least a bit less off topic,
some of the problems underlying the difficulties of designing good
voting systems could equally well apply to the case of selecting one
of three job applicants A, B, and C, where after a careful evaluation
one finds that A is better than B, B better than C, and C better than
A - if what you're doing is ranking them according to at least three
different criteria, then just counting by how many criteria each
applicant is ranked higher than each of the others. Say, by their
intelligence (to the extent that you can gauge it) you rank them
A-B-C; but by their knowledge you rank them B-C-A, while by their
reliability you rank them C-A-B. Now A beats B in two of the three
criteria, while B beats C and C beats A in the same way.
--
* Harald Hanche-Olsen <URL:http://www.math.ntnu.no/~hanche/>
- Yes it works in practice - but does it work in theory?
* Harald Hanche-Olsen
| This is probably equivalent to a system in which every voter has a fixed
| number of (positive) points to distribute among candidates as he wishes:
| Just add a constant to the points awarded by each voter in your system to
| see this.
Not quite. There are two differences. The first is what the absence of
a vote means. In a system with negative scores, absence means zero. In
a system with a skewed scale, absence of votes is generally not tolerated
and voters have to give scores to every candidate. This is known to fail
miserably because once you ask people to rate things below the "don't
care" limit, their score values are completely random. It is therefore
important to let voters decide not to score a particular candidate, and
that the system give such a zero value, which is the second difference.
The Borda system, which gives more points to the most values and zero to
the least valued specifically requires that each candidate gets scored.
| Such systems are prone the same kind of paradoxes that plague all the
| more conventional systems in existence.
The no-vote-means-zero-score rule removes a significant number of
problems, but not all.
| Anyway, I think a better way to achieve what you want (though not paradox
| free - no fair voting system can be paradox free) is the single
| transferrable vote. Each voter ranks all candidates.
This is a huge problem. Voters must be allowed /not to care/ about the
relative ranking of candidates on whom they have no opinion. Forcing
voters to care about these candidates is known to produce lots of noise.
Is this article you have read? http://www.sciencenews.org/20021102/bob8.asp
--
Erik Naggum, Oslo, Norway
Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
+ Erik Naggum <····@naggum.no>:
| * Harald Hanche-Olsen
| | This is probably equivalent to a system in which every voter has a
| | fixed number of (positive) points to distribute among candidates
| | as he wishes: Just add a constant to the points awarded by each
| | voter in your system to see this.
|
| Not quite. There are two differences. The first is what the
| absence of a vote means. In a system with negative scores,
| absence means zero. In a system with a skewed scale, absence of
| votes is generally not tolerated and voters have to give scores to
| every candidate. This is known to fail miserably because once you
| ask people to rate things below the "don't care" limit, their
| score values are completely random.
That's a good point, but I am not sure if the difference is merely
psychological or substantial. I guess it depends on the exact rules
by which scores may be assigned by the voter.
| | Such systems are prone the same kind of paradoxes that plague all the
| | more conventional systems in existence.
|
| The no-vote-means-zero-score rule removes a significant number of
| problems, but not all.
Is this rule (or family of rules?) known by any other names? Do you
know where I can learn more about them?
| | Anyway, I think a better way to achieve what you want (though not
| | paradox free - no fair voting system can be paradox free) is the
| | single transferrable vote. Each voter ranks all candidates.
|
| This is a huge problem. Voters must be allowed /not to care/
| about the relative ranking of candidates on whom they have no
| opinion. Forcing voters to care about these candidates is known
| to produce lots of noise.
I have been worried about this problem. I wonder what would happen to
this system if you simply allowed voters to not rank all candidates?
A ballot on which all the ranked candidates have been removed from the
race might either not be counted anymore, or it could be counted as
1/Nth of a vote for each remaining candidate.
| Is this article you have read?
| http://www.sciencenews.org/20021102/bob8.asp
No, but thanks for the reference. My reading includes a small paper
by Aanund Hylland, "Proportional representation without party lists"
from 1991, and the book "Chaotic elections: A mathematician looks at
voting" by Donald G. Saari, 2000; ISBN 0821828479. Saari is clearly a
big fan of the Borda count, as it is the unique system that minimizes
paradoxes in some sense. But his book does not compare the Borda
count with the single transferable vote (aka instant runoff). From
the reference you just provided, it seems he will do just that in an
upcoming paper. I'll look at it as soon as I can.
My interest in this is not purely theoretical: At my university all
elections are run on a point based system, whereby ranks assigned by
voters are translated into weights 1, 1/3, 1/5, 1/7, ... This seems
to me just a minor perturbation of a simple plurality system. After
each election there is a lot of grumbling by the losers, by nobody
ever suggests a different system. I'd like to do something about
that.
We could go on to talk about this at length, but this is c.l.LISP
after all, so maybe this is a good place to stop (I'll let you have
the last word if you wish).
--
* Harald Hanche-Olsen <URL:http://www.math.ntnu.no/~hanche/>
- Yes it works in practice - but does it work in theory?
"Kenny Tilton" <·······@nyc.rr.com> wrote in message
·····················@nyc.rr.com...
>
> This was a new one on me, and was only one bloke's input. I scoff at
> certification for programmers, but if it would be easy to implement and
> if folks in tall buildings care (two big ifs) why not?
>
Well, I (and probably many people smarter and more talented than me)
object to these sorts of tests on principle. They tend to measure whether
programmer X can do cookie-cutter task Y using language Z (and subsidiary
APIs),
not whether programmer X is any good at, say, problem solving or even, say,
programming.
Most corporations that judge competence and make hiring decisions based
on some third-party jokey-joke test are not the kinds of places that smart
and talented
people are going to enjoy working.
Certifications for other languages have reduced programming from a
profession
to a job, and changed most 'programmers' from engineers to mere technicians.
I would guess that many of the best programmers that choose Common Lisp
would
refuse, on priciple, to go down that road. So what would the test
accomplish?
I envision it inviting hordes of loathsome know-nothing jobbers into the
field,
and sullying the reputations of all serious programmers, making it even more
dificult to find a job, for lack of some idiotic certification.
Screw MegaCorp, I say, let them choke on their own vomit.
-James Russell
"J L Russell" <·········@alum.mit.edu> wrote in message ····························@bgtnsc05-news.ops.worldnet.att.net...
> loathsome
Speaking of loathsome, I apologize for the atrocious line breaking.
-James Russell
J L Russell wrote:
> "Kenny Tilton" <·······@nyc.rr.com> wrote in message
> ·····················@nyc.rr.com...
>
>>This was a new one on me, and was only one bloke's input. I scoff at
>>certification for programmers, but if it would be easy to implement and
>>if folks in tall buildings care (two big ifs) why not?
>>
..snip...
> Screw MegaCorp, I say, let them choke on their own vomit.
The MegaCorp exec was attending a rare Lisper conference and hearing
endless discussion of how CL could expand its adoption. In that context
he suggested certification would grease the skids for adoption by
fortune 500. I agree with you about certficates, and told the guy as
much. he pointed out to me that he was just saying this might be the
price of the f5 acceptance we were after, so it does not matter what we
think of the process.
--
kenny tilton
clinisys, inc
---------------------------------------------------------------
""Well, I've wrestled with reality for thirty-five years, Doctor,
and I'm happy to state I finally won out over it.""
Elwood P. Dowd
Kenny Tilton <·······@nyc.rr.com> wrote in
·····················@nyc.rr.com:
>
>
> J L Russell wrote:
>> "Kenny Tilton" <·······@nyc.rr.com> wrote in message
>> ·····················@nyc.rr.com...
>>
>>>This was a new one on me, and was only one bloke's input. I scoff at
>>>certification for programmers, but if it would be easy to implement
>>>and if folks in tall buildings care (two big ifs) why not?
>>>
> ..snip...
>
>> Screw MegaCorp, I say, let them choke on their own vomit.
>
> The MegaCorp exec was attending a rare Lisper conference and hearing
> endless discussion of how CL could expand its adoption. In that
> context he suggested certification would grease the skids for adoption
> by fortune 500. I agree with you about certficates, and told the guy
> as much. he pointed out to me that he was just saying this might be
> the price of the f5 acceptance we were after, so it does not matter
> what we think of the process.
>
F500 companys employ a lot of programmers, how do they transition them
to lisp? Also the learning curve to be a productive lisp programmer
apears to be about 2 years, from comments I have read here, and the avg
IT job lasts about 2 years in the US. So the IT manager can not justify
training or growing them to his boss. It is like giving the person a
masters degree and then he leave, not good.
And as Erik pointed out the idea of a Java type cert is a generaly bad
idea.
marc
Marc Spitzer <········@optonline.net> wrote in
·········································@167.206.3.3:
>
> F500 companys employ a lot of programmers, how do they transition them
> to lisp? Also the learning curve to be a productive lisp programmer
> apears to be about 2 years, from comments I have read here, and the avg
> IT job lasts about 2 years in the US. So the IT manager can not justify
> training or growing them to his boss. It is like giving the person a
> masters degree and then he leave, not good.
>
> And as Erik pointed out the idea of a Java type cert is a generaly bad
> idea.
>
> marc
>
Between this post and my others on this topic I am arguing both sides of
the issue. Hopefully a working solution is some where in the middle.
marc
On Sun, 03 Nov 2002 05:49:28 GMT, Kenny Tilton <·······@nyc.rr.com>
wrote:
>I agree with you about certficates, and told the guy as
>much. he pointed out to me that he was just saying this might be the
>price of the f5 acceptance we were after, so it does not matter what we
>think of the process.
I'm a programmer not a manager, so my opinion may well be worth just
what you're paying me for it, but - I think the guy is mistaken.
Certification is a consequence, not a cause, of Java's popularity.
If you want a job as a programmer in a Fortune 500 company, the
effective way to get it is to learn Java. If you want to program in
Lisp in such a company, get in as a Java programmer, work there until
you're on the inside track and you've some political power in the
place, then quietly start introducing Lisp in the contexts where it's
most appropriate and will do least to disrupt the existing
infrastructure.
If that's not what you're looking for, you probably don't need a
certificate.
--
"Mercy to the guilty is treachery to the innocent."
Remove killer rodent from address to reply.
http://www.esatclear.ie/~rwallace
On Sat, 02 Nov 2002 17:05:42 GMT, Kenny Tilton <·······@nyc.rr.com> wrote:
> There was a fellow from MegaCorp (manager, not a programmer) who did the
> conference to check out Lisp for a new $5m project. It's a C++ shop and
[...]
> One specific: he said the ALU should cook up a certification exam. He
> understood that certification might be a joke in our domain, but that
> nevertheless it was the kind of thing that would make Lisp look more
> respectable to MegaCorp types.
Some companies offer certification exams (Microsoft? Oracle?). Would it be
acceptable for the MegaCorp manager for a Lisp vendor, such as Franz, to
deliver a modified version of one of its classes that specifically includes
an exam? Maybe such classes already include a final exam, I don't know much
about this.
Paolo
--
EncyCMUCLopedia * Extensive collection of CMU Common Lisp documentation
http://www.paoloamoroso.it/ency/README