From: Tentacle of Igor Chudov
Subject: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies
Date:
Message-ID: <45h5gt$lom@taco.cc.ncsu.edu>
Robert Martin (·······@rcmcon.com) wrote:
: >(This implies that object-oriented technology in the broader sense can
: >impart a sense of joy and accomplishment, which beats bone-head
: >programming in C, C++, or SmallTalk.)
: I understand the "sense of joy", but it is not isolated to OO. I have
: experienced this "joy" when using other methods too. I have
: experienced this "joy" when using no method at all. I have even
: experienced this "joy" when writing bone-head programs in C, C++ and
: Smalltalk. That "joy" is the reason that I am a software engineer.
: (Would that that were true for more software engineers.) I love
: designing and writing sofware. It is a good thing people pay me for
: it, because if they didn't, I would pay them.
Actually, in many ways programming is close to poetry. Just as in
poetry, good programmers feel joy and inspiration solving interesting
problems. Reading a good program is no more boring than reading a
good poem. Small programs can be just as funny as poetic jokes. I
myself experience great pleasure writing code that I consider good. A
great emotional attachment to programming languages and techniques
cannot be explained by rational reasons.
Just as with poetry, there are true geniuses like Knuth, Stallman
or Ritchie, and programming has its share of careless people,
graphomaniacs and crooks. People who write programs for salary's sake
only are just as universally despised as greedy poets. There are even
"schools" in programming, mainstream and non-mainstream methods.
There is even programming decadance (obfuscated C).
Good programming solutions live forever and are infinitely reused.
Programming is poetry for the right brain hemisphere. It is just
as inspirational - remember programmers who work nights long. It
spills it influence to other areas, such as mathematics. As Misha
Verbitsky suggested, some works of programmers (GNU Manifesto)
extend even to the heavens of philosophy.
Unfortunately, left-hemisphere ("normal") poets and programmers
have hard times understanding and recognizing each other.
- Igor.
In article <··········@taco.cc.ncsu.edu>, ·······@galstar.com (Igor
Chudov) wrote:
[SNIP!]
> Programming is poetry for the right brain hemisphere. It is just
> as inspirational - remember programmers who work nights long. It
> spills it influence to other areas, such as mathematics. As Misha
> Verbitsky suggested, some works of programmers (GNU Manifesto)
> extend even to the heavens of philosophy.
Yes, I find the GNU manifesto rather beautiful as well. It infects friends
of mine who care about programming with a shared sense of the possiblities
of community among programmers. I remember it every time I feel the urge
to hoard information rather than share it, and make the right choice.
Can we nominate Richard Stallman for a Nobel prize?
Chris
-------------------------------------
-- Chris Morgan, BAeSEMA Limited
-- ············@baesema.co.uk
-------------------------------------
-- Team Ada
-------------------------------------
From: Tentacle of Igor Chudov
Subject: Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies
Date:
Message-ID: <45okg3$st@taco.cc.ncsu.edu>
Chris Morgan (············@baesema.co.uk) wrote:
: In article <··········@taco.cc.ncsu.edu>, ·······@galstar.com (Igor
: Chudov) wrote:
: [SNIP!]
: > Programming is poetry for the right brain hemisphere. It is just
: > as inspirational - remember programmers who work nights long. It
: > spills it influence to other areas, such as mathematics. As Misha
: > Verbitsky suggested, some works of programmers (GNU Manifesto)
: > extend even to the heavens of philosophy.
: Yes, I find the GNU manifesto rather beautiful as well. It infects friends
: of mine who care about programming with a shared sense of the possiblities
: of community among programmers. I remember it every time I feel the urge
: to hoard information rather than share it, and make the right choice.
: Can we nominate Richard Stallman for a Nobel prize?
I'd be for it, wholeheartedly.
- Igor.
From: Robert Dewar
Subject: Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies
Date:
Message-ID: <dewar.813758236@schonberg>
Well I am not sure that any Nobel prize really applies to the cause of
free software. Nobel was rather specific on the fields to which the
prize should be given (there is for example no Nobel prize in mathematics,
supposedly because his wife ran off with a mathematician)
Note that Stallman did receive a McArthur "genius" award, which was a
well-deserved acknowledgement of his contributions.
From: Simon Brooke
Subject: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <4662m6$ng2@caleddon.dircon.co.uk>
······@cec-services.com (Colin James III (The Rt Rev'd)) wrote:
>| In article <············@wildcard.demon.co.uk>,
>| Cyber Surfer <············@wildcard.demon.co.uk> wrote:
>| >> Can we nominate Richard Stallman for a Nobel prize?
>What on earth does this have to do with comp.lang.eiffel, much less
>comp.lang.ada or comp.lang.lisp.
It has this to do: all of us, whether we write eiffel, ada, or LISP, are
programmers first. There are some people -- a very few -- who through their
contribution have changed the nature of the task and thus of our lives.
To name my particular heroes (one could name others)
Alan Turing set the intellectual framework for programming;
Donald Knuth contributed greatly to the aesthetic framework;
Richard Stallman has contributed greatly to the ethical framework.
In fact is is Stallman who has essentially created the claim that programming
is ethically different from other creative activities, and that claim has
affected the way many of us view the work we do. Whether the claim was true
when Stallman first made it, it is true now, and that's largely because of
Stallman's work. He and his followers have made powerful software systems
available to anyone who can afford the most basic modern hardware. In the
long run this will have a powerfully subversive and democratising effect not
just on the software community but on society as a whole. That's a big
achievement.
--
------- ·····@rheged.dircon.co.uk (Simon Brooke)
'Victories are not solutions.'
;; John Hume, Northern Irish politician, on Radio Scotland 1/2/95
[Colin James III]
| The first two first rate scholars; Don Knuth is a first rate
| educator.
|
| Stallman is not a scholar, or an educator. Stallman is an avowed
| humanist, placing man at the center of the universe, not God.
a religious fanatic at work. how quaint.
| Stallman has had absolutely no effect whatsoever on the development of
| Eiffel, as far as I know, and no Eiffel compiler vendor has bought into
| the lame gnu license copywrong.
this may be why Eiffel is not much used.
| Your argument is invalid, and your ethics reek of subterfuge.
before anyone replies to "the retired reverend", note that his attitude to
responding to those who reply to him is as follows:
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Colin James III (The Rt Rev'd) does not read electronic mail which is
unsolicited, as was the post above, but may be reached by US Mail at:
CEC Services
2080 Kipling Street
Lakewood, Colorado 80215-1502 U S A
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
ignore him. there are always people who will work against anything great.
typically, they hide behind some religion to denigrate good men to please
their own fantasies.
RMS has certainly changed my mind over the past decade, from ardent
believer in the right to control software, but this is like refusing
students of the theater to read plays unless they pay for the right to
perform it, to ardent believer in the benefits of reading source code.
like in all other written art, one learns by studying the masters. RMS has
made this possible and now also popular in the art of software. if
anything, the lasting effect of his work will be improved skills among
programmers, a true historic effect on the art. considering the quality of
commercial software, we should be thankful that someone is workign to give
students of programming access to the great lore of their predecessors,
instead of reinventing wheels over and over again.
I'd nominate RMS for a Nobel Prize in Literature, for bringing the art of
writing computer software into the realm of the literate.
#<Erik 3023193775>
--
a good picture may well be worth a thousand words, but on the WWW,
even bad imagemaps cost tens of thousands of words.
From: Erik Naggum <Erik at NAGGUM.NO>
[re: sharing sources, etc]
RMS has made this possible and now also popular in the art of software.
I'm sorry, but RMS has done no such thing. The sharing of source code
has always been popular amongst Real Computers. All RMS has done is
encourage it in the eunuchs world. In fact, it used to be normal that
when you bought a computer, you got the monitor sources with it.
RMS was an extremely brilliant person. His contribution to the computer
world -before- 1980 was exceptional. But somewhere between 1980-83
something seems to have happened to him: he was forced to cram a 36-bit
brain into a 32-bit world load; it seems to be turning him into cream
cheese.
RMS' ITS EMACS was amazing and a Great Leap Forward; RMS' GNU Emacs
pales in comparison to what other people have done--Symbolics Zmacs,
Twenex EMACS, Multics Emacs, etc. This I believe is primarily because
of the difficulty of writing decent software on a eunuchs platform. I
admit GNU Emacs comes closer to winning than anything else that exists
on eunuchs, but in the end it is dead on the game grid.
I think RMS' problem -is- eunuchs, in fact: the decay of his brain is
well coordinated with his eunuchification.
I disagree that RMS should have a nobel prize--it wouldn't do him any
good. What I think would do him the most good is for someone to give
him a KL10 and a TV terminal, and incinerate all his current losing
machines.
You know, eunuchs really is a damn good name for such a dickless
operating system.
[Erik Naggum]
| some of us think eunuchs is _better_ than the (d)ossified crap that is
| sold to the masses.
[Daniel Finster]
| I wouldn't know. I havn't used an MS-DOS system in over 3 years.
| MS-DOS is hideous, but there are ways in which eunuchs is worse.
you missed the point. by emphasising "better", I intended to prepare for
the point that was coming up later, about marketing. we use "better" tools,
not "good" tools. at the very least, we focus on the better sides, not on
the worst possible aspects. otherwise, going nuts fast is inevitable.
| He's still a flake though. Here's a piece of some mail I got in
| response to my original post:
really? we discussing the psychology of RMS, now? you're taking us on a
field trip of your strongest dislikes, aren't you, Daniel? what for?
| I just wanted you to know that you are not at all off-base in
| your assessment. Surely, a huge load of weird derelict disciples
| of Stallman will send you reams of hate mail accusing you of hate
| crimes and all sorts of lunatic ravings. Just ignore them. Most
| of the flakes are also homeless bums who call themselves ``GNU
| volunteers''.
I wonder what's wrong with whoever wrote this. I've seen a lot of people
express the most intense hatred for RMS, with an accompanying deep interest
in vilifying those who work on GNU projects or with him. frankly, whatever
it is that might be true in what they say, why should anybody listen to
them when _they_ don't want others to listen to someone that _they_ think
is mad? it doesn't make much sense, does it?
I think I shall never see... someone able to criticize RMS or GNU without
going into patently silly attacks of rampant emotionalism. just because
one thinks something that someone does is good, deserves applause, and
should be widely recognized, doesn't mean that one is in love with the guy,
is willing to swallow his every utterance, is a "weird derelict disciple"
(nor an ordinary disciple), nor will defend him on every count, but it
does, of course, mean that those who attack him on the basis of irrelevant
feelings or issues should be countered simply because they do not present
the whole truth, only the part of it that they self-servingly want to be
perceived as the whole truth in order to be relieved of the duty to present
actual arguments. a clear case of "argumentum ad hominem".
_any_ sense of fairness and justice demands that both sides be heard, and
the kind of one-sided prejudice that we see in attacking and judging RMS
can only be to the discredit of the attackers.
I mean, Louis Farrakhan managed to lead this half-million man march, and
although he is the most despicable black man that walks the surface of the
North American land, people _still_ could discuss the purpose of the march
without getting all worked up and getting fanatically myopic about the
issue of his racism, anti-semittism, and sexism. how? the man is clearly
sick and sick men who manage to command half a million men is a clear and
present danger to any society. but it's RMS that needs to be attacked.
pity the small minds that can see no further than their nose.
furthermore, normal people lead normal lives. those individuals who cross
the prevalent views in society will be branded as mad and shunned by their
contemporaries, but may be hailed as forerunners by a future society and by
those who are less fearful of change and novelty. being a pioneer in any
field is always a sign of unwillingness to succumb to what currently passes
for "normalcy". given this, what point is there in ranting and raving with
those silly "he's a basket case" labels? it can but unite those who _are_
weird derelict disciples. it can also only nuture any latent or present
paranoia: it is a matter of fact that people _are_ out to get RMS, right?
I think "irrational fear of RMS" should be a recognized phobia.
#<Erik 3023632684>
--
a good picture may well be worth a thousand words, but on the WWW,
even bad imagemaps cost tens of thousands of words.
In article <················@naggum.no>, Erik Naggum <····@naggum.no> wrote:
[SNIP]
Hello,
I seem to remember this thread starting off with praise for the GNU
manifesto and leading to discussions of GNU software and a suggestion of a
Nobel prize nomination for RMS. Now I pick up the thread again to find
it's been dragged down to personal attacks. Erik you're dealing with
someone with a major chip on their shoulder, how do you remain calm?
All that stuff about RMS as a student, RMS in 1982 is totally
irrelevant. GNU project is a success, I and many other people use their
software every day. GCC gave GNAT a headstart without a doubt. Emacs is a
powerful multi-function editor/development environment if that's what you
want (some people prefer pretty but limited editors (textedit) or minimal
Unix editors (vi)). These are all facts which influence me in my view
(which I repeat) that the GNU manifesto is rather beautiful. When I read
about the activities of Microsoft which appears to me to be attempting the
biggest vendor lock-in in history, I start to feel paranoid myself...
Anyway, having very briefly corresponded with the great man himself, I
would like to report his message that the best profit margin is obtained
by the FSF from CD-ROMS, so although buying manuals helps, buying the
CD-ROMS helps even more.
Chris
-------------------------------------
-- Chris Morgan, BAeSEMA Limited
-- ············@baesema.co.uk
-------------------------------------
-- Team Ada
-------------------------------------
From: Erik Naggum <ERIK at NAGGUM.NO>
we use "better" tools, not "good" tools. at the very least, we
focus on the better sides, not on the worst possible aspects.
otherwise, going nuts fast is inevitable.
That may be true for you; I use -good- tools, like Symbolics Genera.
I choose to avoid tools that are anything less than the best.
I wonder what's wrong with whoever wrote this.
I wonder what's wrong with the cretins that want to give him a nobel
prize.
I've seen a lot of people express the most intense hatred for RMS,
with an accompanying deep interest in vilifying those who work on
GNU projects or with him.
I said it in my original post and I'll say it again: RMS -used to be-
(before 1980 or so) an -extremely brilliant person-, who made a lot of
important contributions to the design of tools I use every day. I do
not vilify or despise the people who write GNU software. GNU software
and GNU.AI.MIT.EDU machines have been very valuable to me at times
that I've been forced to wade through the slime of eunuchs. I don't
even really -hate- RMS; I just think he has gone totally off into
lossage, and I really get sick of people thinking he's some kind of
god, or genius, or messiah, or whatever.
I'm not opposed to the concepts of free flow of information, and
sharing programs with others. I -AM- opposed to the particular
communist overtones of RMS' philosophy, expressed in the GNU
Manifesto.
does, of course, mean that those who attack him on the basis of
irrelevant feelings or issues should be countered simply because
they do not present the whole truth, only the part of it that they
self-servingly want to be perceived as the whole truth in order to
be relieved of the duty to present actual arguments.
The same is true of those that are seriously enthralled by the guy:
they do not present the whole truth, only the part of it that they
self-servingly want to be perceived as the whole truth. They do not
present actual arguments, reverting instead to "RMS is such a cool
stud, and he provides all this nifty software for free, so how dare
you possibly say anything bad about him?"
It goes both ways, sir. But, in the spirit of fairness, I am trying
to present a more on-track viewpoint. You have to realize that I'm
not slamming the guy.
I think he needs to be dumped in the cold-load stream with a bar of
soap. Leave him there for a few hours to clean up, cut his frigging
jesus beard and hippie hair off, and lock him out of the AI Lab for a
few months or so to let people disinfect the building.
Then maybe they can let him back in, after they have replaced his
eunuchs workstation with a SHOWER.
In article <··········@SDF.LONESTAR.ORG>,
Daniel Finster <DF at SDF.LONESTAR.ORG> wrote:
>I'm not opposed to the concepts of free flow of information, and
>sharing programs with others. I -AM- opposed to the particular
>communist overtones of RMS' philosophy, expressed in the GNU
>Manifesto.
I bet you didn't know that RMS possesses a piece of paper with the
names of forty-seven card-carrying atheist programmers written on
it.
>I think he needs to be dumped in the cold-load stream with a bar of
>soap. Leave him there for a few hours to clean up, cut his frigging
>jesus beard and hippie hair off, and lock him out of the AI Lab for a
>few months or so to let people disinfect the building.
Yup. I bet you didn't know that.
--
By sending unsolicited commercially-oriented e-mail to this address, the
sender agrees to pay a $100 flat fee to the recipient for proofreading
services.
In Article <··········@SDF.LONESTAR.ORG>
Daniel Finster <DF at SDF.LONESTAR.ORG> writes:
> From: Erik Naggum <ERIK at NAGGUM.NO>
>
> we use "better" tools, not "good" tools. at the very least, we
> focus on the better sides, not on the worst possible aspects.
> otherwise, going nuts fast is inevitable.
>
>That may be true for you; I use -good- tools, like Symbolics Genera.
>I choose to avoid tools that are anything less than the best.
>
(deletia, deletia, dementia)
>I'm not opposed to the concepts of free flow of information, and
>sharing programs with others. I -AM- opposed to the particular
>communist overtones of RMS' philosophy, expressed in the GNU
>Manifesto.
(more)
>It goes both ways, sir. But, in the spirit of fairness, I am trying
>to present a more on-track viewpoint. You have to realize that I'm
>not slamming the guy.
Sir, you printed a letter detailing RMS's "phobias," including him
"being afraid to go out of his office." Funny, he seems fearless enough to go
to confrences, scifi cons, and resturants. What's up with that? No mention of
*why* you think the GNU Manifesto was "communist," just an all-out slam on the
guy.
>I think he needs to be dumped in the cold-load stream with a bar of
>soap. Leave him there for a few hours to clean up, cut his frigging
>jesus beard and hippie hair off, and lock him out of the AI Lab for a
>few months or so to let people disinfect the building.
Oh gawd, not *long hair*. F*** you sir, and I mean it in the nicest way
possible.
sigh,
just me.
Please get this thread out of comp.lang.lisp. We have better things
to do than slam or defend RMS.
Followups have been set appropriately.
-- Chuck
--
Chuck Fry Work: ······@ptolemy.arc.nasa.gov Play: ······@rahul.net
QuickMail SUCKS! It's slow, verbose, and non-standard.
I alone am responsible for the contents of this post.
Keep NASA and Caelum Research out of this.
[Daniel Finster]
| That may be true for you; I use -good- tools, like Symbolics Genera.
| I choose to avoid tools that are anything less than the best.
that's good for you. however, you again miss the point. the resources
available to any one person are limited, so most people go for the _better_
of the choices available with those resources, instead of the best choice
in a minor area to forfeit the rest. few can afford to refuse to work with
tools and under conditions that they consider inferior. I do, and this
limits my options, as I'm sure Symbolics Genera limits yours (e.g., I rent
a small apartment instead of owning a condo or house like most of my
classmates do these days, I don't have a car, and I neither smoke nor
drink, and so don't need to make more than about half of what passes for
"normal living standards", much to the dismay of the IRS (equivalent), and
thus can work almost solely on projects I consider "worth doing"). some
will find such limits too tight for their comfort, and opt for a wider set
of choices, albeit of less quality.
I think you need to realize that yours is not the only valid choice (of
what constutues "good") in this situation. I think that peace consists of
universal agreement on what are the _wrong_ choices and respect for those
that choose among those that are _not_ wrong. anything better than that is
very hard to obtain, perhaps impossible, in a society of any significant
number of people.
| I wonder what's wrong with the cretins that want to give him a nobel
| prize.
and why did they have to be _cretins_? I think I missed that argument.
| The same is true of those that are seriously enthralled by the guy:
| they do not present the whole truth, only the part of it that they
| self-servingly want to be perceived as the whole truth. They do not
| present actual arguments, reverting instead to "RMS is such a cool
| stud, and he provides all this nifty software for free, so how dare
| you possibly say anything bad about him?"
how many people of this kind do you actually know, Daniel? you attribute
to people opinions and motives they simply do not hold. this is far worse
than missing an occasional shower in my mind.
I _could_ interpret your previous comment as saying that you wonder what's
wrong only with those cretins that happen to join the crowd of brilliant
people that want to give RMS a Nobel prize, but that is not very likely.
the same applies to regarding the above statement as being a depiction of a
small group of lunatics who just happen to be "lunatics for RMS" -- you're
tarring a community of excellent software authors with a mighty broad
brush, and this is so incredibly unfair and unjust that I wonder how you
can live yourself. are you really prepared to face all the people you have
vilified randomly and repeat your statements to them individually?
| But, in the spirit of fairness, I am trying to present a more on-track
| viewpoint. You have to realize that I'm not slamming the guy.
are the following comments of yours "more on-track" and "not slamming"?
| I think he needs to be dumped in the cold-load stream with a bar of
| soap. Leave him there for a few hours to clean up, cut his frigging
| jesus beard and hippie hair off, and lock him out of the AI Lab for a
| few months or so to let people disinfect the building.
|
| Then maybe they can let him back in, after they have replaced his
| eunuchs workstation with a SHOWER.
I'm glad to see that the most important, most "on-track", issue with RMS is
his personal hygiene. that would be somewhat like someone having read my
works on SGML over the last five years and come and visit me and exclaim
"there are _cat_hairs_ everywhere, your whole place smells cat food, your
clothes are full of tiny holes, and your hands have scars from playing with
the furry devils; how can we trust what you have written about SGML?"
I take it that you see the obvious flaw in this "line of reasoning".
I think we have converged on maximal divergence in this thread, Daniel.
there really isn't much more to say. I also don't think you can make the
Nobel committee deliver the Nobel prize in coupons redeemable only at
beauty salons for hackers.
#<Erik 3024513763>
--
a good picture may well be worth a thousand words, but on the WWW,
even bad imagemaps cost tens of thousands of words.
From: Erik Naggum <ERIK at NAGGUM.NO>
and why did they have to be _cretins_? I think I missed that argument.
I think the entire idea of giving RMS (or -ANY- programmer) a Nobel
Prize is the most completely asinine suggestion I have ever seen on
this newsgroup. Much worse even than the Lisp->Perl translator
suggestion.
In other words, it totally annoys me. This being Usenet, the obvious
thing to do is flame-bait until my opposition decides that I'm as
stupid as I've decided that they are.
In reality, I really don't care about RMS. I don't ever think about
him, and he has no place in my world. Much as it may surprise you, I
don't typically go around just badmouthing him to random strangers
(this being a special case). I don't hate him (though I don't really
like him either; maybe a `neutral' feeling?)
I hope that those of you out there with a sense of humor got some
laughs out of all this. I certainly had a lot of fun writing my
posts.
are the following comments of yours "more on-track" and "not slamming"?
Of course not. Thats what makes it so funny.
In article <··········@SDF.LONESTAR.ORG>,
Daniel Finster <DF at SDF.LONESTAR.ORG> wrote:
>In reality, I really don't care about RMS.
I.e. you are trolling.
>I hope that those of you out there with a sense of humor got some
>laughs out of all this. I certainly had a lot of fun writing my
>posts.
You're welcome.
P.S. Fix your fucking newsreader already, cretin.
--
By sending unsolicited commercially-oriented e-mail to this address, the
sender agrees to pay a $100 flat fee to the recipient for proofreading
services.
In article <··········@news4.digex.net> ···@access2.digex.net (Ell) writes:
>Daniel Finster (DFatSDF.LONESTAR.ORG) wrote:
>
>I support giving RMS the Nobel, or whatever the highest award is for the
>kind of contributions he has made. To coordinate and contribute to the
>cornucopia of FREE software, of such importance, from GNU is worthy of
>our highest honors.
That is a bit too far. The Nobel prize is presumably for people who
have made great contributions to humanity. However RMS definitely does
deserve a high honor in the Computer World. Go no further than Linux
to see how his vision of free software has produced what many people
regard as possibly the best OS that was ever written for the IBM PC.
--Jag
--
/\/\ |Jagadeesh K. Venugopal, ·····@ctp.com |All opinions expressed here |
/ /_.\|Cambridge Technology Partners, Inc. |are my personal opinions only |
\ /./|304 Vassar St. Cambridge, MA 02139 |and not those of my employer. |
\/\/ |Voice Mail: (617) 374-2028. +______________________________+
In article <··········@Starbase.NeoSoft.COM>, ······@Starbase.NeoSoft.COM
(Cameron Laird) wrote:
> There's no need to presume. <URL:http://www.nobel.ki.se/cgi-bin/
> uncgi/nobel?to=17;;&lng=0&bkp=4&ctr=1&picb=0&gfx=1> excerpts and
> translates Nobel's will:
>
> ... annually distributed in the form of prizes
> to those who, during the preceding year, shall
> have conferred the greatest benefit on mankind.
> The said interest shall be divided into five
> equal parts, which shall be apportioned as fol-
> lows: one part to the person who shall have made
> the most important discovery or invention within
> the field of physics; one part to the person who
> shall have made the most important chemical
> discovery or improvement; one part to the person
> who shall have made the most important discovery
> within the domain of physiology or medicine; one
> part to; and one part to the per-
> son who shall have done the most or the best work
> for fraternity between nations, for the abolition
> or reduction of standing armies and for the hold-
> ing and promotion of peace congresses.
>
> 'Doesn't seem to me that RMS qualifies. Why was this in dispute?
Looks to me like his essays on why software source should be a free good
fit _exactly_ into the category "the person who shall have produced in the
field of literature the most outstanding work of an idealistic tendency"
[where in fact "work" usually ends up being interpreted as "body of
work"]. Now all that is needed is for the nomination process to begin.
Thank you for the quote.
--
Xanthian. | "..want the consequences of what you want.." |
Kent, the man from xanth. | Neil A. Maxwell, LDS Apostle |
Kent Paul Dolan ------------------------------------------------
········@{well,qualcomm}.com Jobhunting? Check www.qualcomm.com!
In article <··········@SDF.LONESTAR.ORG>,
Daniel Finster <DF at SDF.LONESTAR.ORG> wrote:
>I think the entire idea of giving RMS (or -ANY- programmer) a Nobel
>Prize is the most completely asinine suggestion I have ever seen on
>this newsgroup. Much worse even than the Lisp->Perl translator
>suggestion.
Does it bother you that Stallman is vivid?
>In other words, it totally annoys me. This being Usenet, the obvious
>thing to do is flame-bait until my opposition decides that I'm as
>stupid as I've decided that they are.
>
>In reality, I really don't care about RMS. I don't ever think about
>him, and he has no place in my world. Much as it may surprise you, I
>don't typically go around just badmouthing him to random strangers
>(this being a special case). I don't hate him (though I don't really
>like him either; maybe a `neutral' feeling?)
I believe you are afraid of him.
[...]
[Yes, it's from Emacs' Doctor mode.]
--
And I will CONTINUE this message - next week!
--
Dan
.... And a nice day to You too, Sir.
--
(Rmz)
Bj\o rn Remseth !Institutt for Informatikk !Net: ···@ifi.uio.no
Phone:+47 22855802!Universitetet i Oslo, Norway !ICBM: N595625E104337
From: Bruce Link
Subject: Re: Ethics in programming; not likel
Date:
Message-ID: <468n7j$sh6@deneb>
Colin James III allegedly wrote with possible deletions:
: Stallman is not a scholar, or an educator. Stallman is an avowed
: humanist, placing man at the center of the universe, not God.
The religious views of Stallman, or Colin have nothing to do with the topics
of these groups, and have no place here.
: Stallman has had absolutely no effect whatsoever on the development of
: Eiffel, as far as I know, and no Eiffel compiler vendor has bought
: into the lame gnu license copywrong.
Why are you discussing Eiffel in your post to comp.software-eng,comp.object,
comp.lang.ada,comp.lang.lisp? You didn't even include comp.lang.eiffel.
: Your argument is invalid, and your ethics reek of subterfuge.
These ad hominem attacks are not worthy of someone who claims religious
righteousness. I can only assume that some nut is impersonating the "Rt.
Rev'd" Colin James III.
: ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
: Colin James III, Principal Scientist ······@cec-services.com
: CEC Services, 2080 Kipling St, Lakewood, CO 80215-1502 USA
: Voice: 303.231.9437; Facsimile: .231.9438; Data: .231.9434
: ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
------------------------------------------------------------------------------
Bruce Link |Team OS/2
···@mda.ca or ·····@wimsey.com |Team Ada
Colin James III (The Rt Rev'd) (······@cec-services.com) wrote:
* The first two first rate scholars; Don Knuth is a first rate
* educator.
* Stallman is not a scholar, or an educator. Stallman is an avowed
* humanist, placing man at the center of the universe, not God.
:)
* Stallman has had absolutely no effect whatsoever on the development of
* Eiffel, as far as I know, and no Eiffel compiler vendor has bought
* into the lame gnu license copywrong.
Which is one of the main reasons why Eiffel has little popularity.
Think about it.
--
- Igor. (My opinions only) http://www.galstar.com/~ichudov/index.html
For public PGP key, finger me or send email with Subject "send pgp key"
You know you have achieved perfection in design, not when you have nothing
more to add, but what you have nothing more to take away.
- Antoine de Saint Exupery.
From: Chris Morgan
Subject: Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <chris.morgan-2010951608180001@baesema4.demon.co.uk>
In article <··········@caleddon.dircon.co.uk>, Simon Brooke
<·····@rheged.dircon.co.uk> wrote:
[SNIP]
programmers first. There are some people -- a very few -- who through their
> contribution have changed the nature of the task and thus of our lives.
>
> To name my particular heroes (one could name others)
> Alan Turing set the intellectual framework for programming;
> Donald Knuth contributed greatly to the aesthetic framework;
> Richard Stallman has contributed greatly to the ethical framework.
[SNIP]
In addition to the GNU manifesto and project (especially emacs and GCC)
which most here know and appreciate, I think his insight into the GNAT
library model (no library, reparse on the fly) was a major breakthrough.
With enough memory around, GNAT demolishes all other known compilers. When
I go back to any old compiler I just don't like it (maybe that's
irrational but it's true).
Anyway, what I believe Richard wants more than gongs is contributions,
which is why I got my company to buy a box full of manuals for GCC and
emacs from them. This helps the FSF and so furthers the GNU ideal. The
fact that they are well-produced and exceptionally well written, and emacs
is boosting productivity all around me is, or course, very welcome as
well. You buy manuals from Sun if you buy their software, so you should
consider buying manuals from FSF if you use their software to ensure it
continues to develop.
e-mail responses are more likely to reach me.
Chris
-------------------------------------
-- Chris Morgan, BAeSEMA Limited
-- ············@baesema.co.uk
-------------------------------------
-- Team Ada
-------------------------------------
From: Jay Martin
Subject: Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <468q9m$c2m@saba.info.ucla.edu>
············@baesema.co.uk (Chris Morgan) writes:
>In article <··········@caleddon.dircon.co.uk>, Simon Brooke
><·····@rheged.dircon.co.uk> wrote:
>[SNIP]
> programmers first. There are some people -- a very few -- who through their
>> contribution have changed the nature of the task and thus of our lives.
>>
>> To name my particular heroes (one could name others)
>> Alan Turing set the intellectual framework for programming;
>> Donald Knuth contributed greatly to the aesthetic framework;
>> Richard Stallman has contributed greatly to the ethical framework.
>[SNIP]
>In addition to the GNU manifesto and project (especially emacs and GCC)
>which most here know and appreciate, I think his insight into the GNAT
>library model (no library, reparse on the fly) was a major breakthrough.
>With enough memory around, GNAT demolishes all other known compilers. When
>I go back to any old compiler I just don't like it (maybe that's
>irrational but it's true).
I think the GNAT library design is trash, not some miracle of Computer
Science. My objection is that it makes a connection between file
names and library units that shouldn't be there. Treating Ada files
like ".h" include files is not my idea of brilliant. There is no
reason in my mind why the classical library approach could not be same
or faster than the ".h" include approach. I am not impressed with GNU
software since it is written in (or compiles) C which in my view just
degrades and stops progress in the programming community. I use GCC
and emacs but I am not really impressed with the design of emacs.
I think Stallmans leftist free-software philosophy/movement is great.
On to Knuth, his contribution to the field of Algorithms and CS Math
is unquestionable. I question more the genis "programming" side.
- His Algorithm books used (bleh) assembly.
- Defended the use of GOTO's in the 70's.
- I am not impressed with "Literate Programming" which basically
just shoved typeset comments in the code. Even worse the system
took over the structuring of programs: subroutines, abstract
data structures, objects, etc and replaced them with some silly
paragraph structure and super macro pre-processor.
I think more of: Ichbiah, Stroustrup and Meyers in the programming area.
Jay
From: Chris Morgan
Subject: Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <chris.morgan-2110951325550001@baesema4.demon.co.uk>
In article <··········@saba.info.ucla.edu>, ·······@cs.ucla.edu (Jay
Martin) wrote:
[SNIP]
> I think the GNAT library design is trash, not some miracle of Computer
> Science. My objection is that it makes a connection between file
> names and library units that shouldn't be there. Treating Ada files
> like ".h" include files is not my idea of brilliant. There is no
> reason in my mind why the classical library approach could not be same
> or faster than the ".h" include approach. I am not impressed with GNU
> software since it is written in (or compiles) C which in my view just
> degrades and stops progress in the programming community. I use GCC
> and emacs but I am not really impressed with the design of emacs.
> I think Stallmans leftist free-software philosophy/movement is great.
[SNIP]
I dont think its a "miracle" either. However it's a big step forward
compared to other Ada compilers that I have used. There is an extremely
important distinction between the GNAT library model and C or C++. For
example, I don't recall having to put hash defines in Ada specs to stop
the compiler doing them twice as you have to in the C++ I use. As to the
filename limitation, well I'm sure someone will eventually do the tool
that hides this from you if you want. I personally don't see a problem.
As for GNU being poor because of C, well that is just prejudice! How
many other people do you think would criticize the writers of an operating
system for using C? The phrase portable assembler describes C perfectly.
Emacs is mostly written in elisp, and now with GNAT you could contribute
software in Ada95 if you so chose.
At least we can agree on the GNU ideal anyway.
Regards,
Chris
-------------------------------------
-- Chris Morgan, BAeSEMA Limited
-- ············@baesema.co.uk
-------------------------------------
-- Team Ada
-------------------------------------
From: ·······@inmind.com
Subject: Re C as portable assemble (was Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies)))
Date:
Message-ID: <46bc28$to7@mujibur.inmind.com>
In <·····························@baesema4.demon.co.uk>, ············@baesema.co.uk (Chris Morgan) writes:
>The phrase portable assembler describes C perfectly.
I have to take exception to this. I have seen people state
this over and over and it is flat wrong. C is no more a portable
assembler than PASCAL is. It is simply richer. It contains
all of the traditional "high level" control structures. It
contains the expected primitive data elements. And it
just happens to contain a few additional operations on
those primitive data elements -- but far from the entire
set that you might expect in an assembler.
What it does NOT contain that assembler languages
contain, as a rule of thumb, are...
1. Registers. The "register" keyword doesn't count because
it isn't really specifying a register. It is just giving the
compiler an optimization hint.
2. Many operators. Where are the arithmetic right shifts?
Certainly (as per the standard) not in C. The vendor may
or may not choose to provide them. Where are the rotates?
Where is the ability to have label variables -- so that you
can directly implement things like the switch control
structure at a low level.
3. Where is the processor information such as carry
flags? Such as result flags from comparisons?
4. Where are the assembler "statements" which correspond
to two or three address expressions. The C expression is
too high level to be considered a candidate because you
cannot control the precise statements generated and their
order of generation.
What would a portable assembler language look like? It would
probably have an infinite number of registers, each of a
specific length. It would have the missing operations. It would
have the carry flag or some other way to directly access the
results of comparisions and shifts/rotates. The carry flag is
esential to many assembler algorithms because it is the basic
mechanism used to (efficiently) combine operations to obtain
similar operations on larger operands. It would have available
the ability to explictly sign or zero extend values. It would
have the X * X -> XX (single precision input to double precision
output using two registers) type of multiplication which is essential
to efficient implementation of unbounded arithmetic precision
packages.
I am interested in defining a portable assembler language that
captures most of the assembler facilities used by application
programming (ignoring the operating system part of the hardware.)
C is a far cry from what is needed. It is simply a high level
language (no less so that PASCAL which nobody has ever
described as a portable assembler language) with a few more
operations than most languages and not as many as some.
Michael Lee Finney
From: Chris Morgan
Subject: Re: Re C as portable assemble (was Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies)))
Date:
Message-ID: <chris.morgan-2210951537320001@baesema4.demon.co.uk>
In article <··········@mujibur.inmind.com>, ·······@inmind.com wrote:
> In <·····························@baesema4.demon.co.uk>,
············@baesema.co.uk (Chris Morgan) writes:
>
> >The phrase portable assembler describes C perfectly.
>
> I have to take exception to this. I have seen people state
> this over and over and it is flat wrong. C is no more a portable
Unix was the first portable operating system because it was rewritten in
C in 1973. WindowsNT is also largely written in C and is hosted on several
leading architectures. This fact is often quoted as one of the reason
these large powerful and successful operating systems exist. No one ever
means to suggest you can do without direct access to assembly language for
certain tasks.
Anyway, if I drop the offending phrase and try to explain what I meant,
C is beneficial to use for some low-level functionality which might
otherwise be expressed in (less-portable) assembler. If a 'real' portable
assembler were made, it might be vastly superior. However in the situation
we have currently, it makes no sense to attack C or GNU software because
Ada is better, since they're not really addressing the same problem.
Cheers,
Chris
-------------------------------------
-- Chris Morgan, BAeSEMA Limited
-- ············@baesema.co.uk
-------------------------------------
-- Team Ada
-------------------------------------
From: Robert Dewar
Subject: Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <dewar.814248052@schonberg>
Jay Martin says:
"or faster than the ".h" include approach. I am not impressed with GNU
software since it is written in (or compiles) C which in my view just"
That's a common misconception with respect to GNAT. It is completely
False. The GNAT front end is written in Ada 95, not in C. It does not
in
any sense compile C or compile into C.
Jay thinks that the library design is trash, but actually his objection seems
rather slight, he just does not like the fact that file names are connected
to unit names.
Fair enough, but let's look at this issue in a little calmer manner. First,
if you want a system with a central library, where there is no connection
between file names and unit names, this is easily achievable with a simple
set of shell scripts. Jean Pierre Rosen has for instance written such a
system as a set of simple REXX scripts for OS/2. Of course if you adopt this
approach then you get all the problems (maybe Jay thinks they are advantages
in his view) of having to follow a specific order of compilation.
It would also be quite easy to optionally have a mechanism of using a
centralized file which did file name translations. Indeed this was part of
the original design.
The advantage is that you break the need for a connection between the names.
The disavdnatgae is that you now have a centralized data structure, which
existing configuration management systems don't know about, and which has
to handle common access from parallel compilations. Also, if this file is
yupdated by compilation, you also have introduced an order of compilation
requirement.
We discussed this a couple of times before, once on the old gnatchat mailing
list and once on comp.lang.ada. The consensus in these discussions was that
it was not worth the trouble, and extra complexity. It is certainly
something that can be revisited.
Right now, my impression is that most GNAT users are either quite happy with
the naming convention (it actually has some definite advantages, you know
immediately what file to look in from the name of the unit). Others find
that gnatchop meets their needs.
But we can certainly revisit the file name mapping option sometime. I
believe that the Intermetrics compiler, while pretty similar to the
GNAT approach in its source orientation, except that it has such a
centralized name translation file.
So far, Jay's note is the only strong vote I have seen for moving in
this direction. For myself, having used many Ada compilers, I much
prefer the GNAT approach. I don't mind having to name my files
canonically, it seems like a good idea in any case, and the advantages
of having absolutely NO centralized data structures, and a compilation
model that is similar to the rest of the world and very light, seems
a real advantage.
So, is Jay a lone voice or a small minority, or is it worth adding
this extra level of optional complexity to GNAT (it would not be a big
implementation effort by any means -- I am more concerned with the
extra complexity of use and description than in the implementation
effort).
From: James W. Bennett
Subject: Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <46avj3$j7@caldonia.tara-lu.com>
In article <···············@schonberg>, ·····@cs.nyu.edu (Robert Dewar) writes:
>Jay thinks that the library design is trash, but actually his objection seems
>rather slight, he just does not like the fact that file names are connected
>to unit names.
[Snip]
>So far, Jay's note is the only strong vote I have seen for moving in
>this direction. For myself, having used many Ada compilers, I much
>prefer the GNAT approach. I don't mind having to name my files
>canonically, it seems like a good idea in any case, and the advantages
>of having absolutely NO centralized data structures, and a compilation
>model that is similar to the rest of the world and very light, seems
>a real advantage.
>
>So, is Jay a lone voice or a small minority, or is it worth adding
>this extra level of optional complexity to GNAT (it would not be a big
>implementation effort by any means -- I am more concerned with the
>extra complexity of use and description than in the implementation
>effort).
>
I wouldn't put it quite as strong as Jay, I don't think the approach is
complete trash. However, after having used many Ada compilers over the
last 10 years, I happen to very much like the DEC approach of library
management.
--
Jim Bennett
····@tara-lu.com
From: Robert Dewar
Subject: Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <dewar.814288174@schonberg>
Jim says:
"I wouldn't put it quite as strong as Jay, I don't think the approach is
complete trash. However, after having used many Ada compilers over the
last 10 years, I happen to very much like the DEC approach of library
management."
One thing to investigate here is what it is that you liked about the
DEC approach. Any or all of it can be simply duplicated with GNAT. As
far as I know, only Jean Pierre Rosen has actually done the work of
duplicating a conventional Ada library system. Since the amount of work
is trivial, the fact that no one has done it says something already!
Things I do NOT like about the conventional library system are:
o Required order of compilation
o One compilation affects others
o Meaning of program depends on order of compilation
o Conventional tools don't understand Ada library
Now if you find these advantages, you can certainly rig up a little script
to give identical effects with GNAT. Here is an outline of how it would work:
Establish a directory to be used as the Ada library. This is a black box as
far as the user of the system is concerned, like a traditional Ada library.
Into this directory will go source, ali, and o files, but the source files
are not the input source files that the programmer deals with.
To compile a new file:
gnatchop the file into a temporary directory (not the Ada library
directory).
In the order that the files were gnatchopped, copy them into the Ada
library directory and compile them.
Note: use the -r switch on the gnatchop so that error messages and the
debugger refer back to the original files.
When you are ready to build a program, use gnatmake to build the main
program from the library directory.
That's all there is to it. This scheme will exactly duplicate the behavior
of a conventional library system, including enforcing an order of compilation,
meaning of program depends on order of compilation, parallel compilation has
to be very careful of the shared data base etc.
I certainly wold not take this approach, because it seems to work hjard to
establish a set of dsadvantages, but one persons' disadvantage is another
persons desirable feature, so if you want this you can have it.
However, when I have talked further with people syaing that they like
for example the DEC library system, it is not so much that they are after
the fundamental semantics of a centralized library system. Instead what
they often are asking for is a nice interface for managing their library.
In normal use of GNAT, your library is your set of sources. You absolutely
need tools for managing this set of sources that are analogous to the kind
of commands used to manage a DEC library.
If that is what you are after, then I think it is better to concentrate
on building these tools with respect to the working set of sources.
Certainly I have dozens of command files (shell scripts) around in my
OS/2 environment to provide this kind of management.
Creating these tools is basically trivial, because there is no mystery
here -- GNAT source files are just ASCII files. For example, you might
want to use CONTINUOUS to manage your source files for you if you
like that
moel.
What we need to build up is a standard set of such tools. One trivial
example that has been around for a while is gnatk8, the little utility
that gives the file name given the unit name. Obviously you need the
inverse of this, that gives the unit name, given the file name. The SGI
version has the latter utility in the distribution, and we should certainly
add it -- it is on the list of things to do! These are of course building
blocks for other scripts.
By the way, the mapping of unit names to file names is entirely concentrated
in one unit of the compiler, Fname, so it is easy to play with alternative
schemes, including for eample a centralized directory.
A nice scheme would be to have a name server that is separate from the
compiler. The default server would just do the alogorithmic translation
that is done now, but alternative servers could provide more general
directory based, or even configuration manager based, approaches to
the mapping from unit names to file names.
This actually seems quite attractive. I wonder how much extra overhead
it would introduce, and how much extra complexity. The server would need
to be some DLL, or DLL equivalent -- interesting, worth looking into!
Anyway, what is helpful in this discussion is to try to be as specific
as you can about what you would like to be able to do, and what features
you would like to see. GNAT is really simply a building block used to
assemble an Ada compilaion system. The approach it takes israther low
level, which has the advantage that it can fit into almost any scheme
that you have in mind.
In some cases, this is achieved simply with a set of scripts (for example,
I never type a gcc command, I use one of my higher level scripts). In
other cases, quite small changes to the compiler may enable solutions
that meet currently unmet needs.
From: Job Honig
Subject: Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <46b8je$bqu@mo6.rc.tudelft.nl>
In article <··········@saba.info.ucla.edu>,
Jay Martin <·······@cs.ucla.edu> wrote:
>
>I think the GNAT library design is trash, not some miracle of Computer
>Science. My objection is that it makes a connection between file
>names and library units that shouldn't be there. Treating Ada files
This has been discussed several times now... If you look at the full set
of Gnat tools, there is no such connection AT ALL. If you use gnatchop,
you may choose the file names you want; gnatchop will generate the
"intermediate" source files that gnat requires.
Job Honig
Delft University of Technology
From: Albert F. Niessner III
Subject: Gnat and Library (was Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies)))
Date:
Message-ID: <AFNIII.95Oct24091056@hades.arl.psu.edu>
In article <···············@schonberg> ·····@cs.nyu.edu (Robert Dewar) writes:
.
.
.
SNIP
Establish a directory to be used as the Ada library. This is a black box as
far as the user of the system is concerned, like a traditional Ada library.
Into this directory will go source, ali, and o files, but the source files
are not the input source files that the programmer deals with.
To compile a new file:
gnatchop the file into a temporary directory (not the Ada library
directory).
In the order that the files were gnatchopped, copy them into the Ada
library directory and compile them.
Note: use the -r switch on the gnatchop so that error messages and the
debugger refer back to the original files.
When you are ready to build a program, use gnatmake to build the main
program from the library directory.
.
.
.
SNIP
This is what I tried to do with the original "librarian" that
I posted a while back. However, the -r on gnatchop does not retain
the path that the source is located in. If I do a
gnatchop -r hello_world I get a
pragma Source_Reference (000001, "hello_world.ada");
This does not have the source path, so if I then go and move the
.ad[bs], .o, and .ali files I loose the correct referencing to
the source -- I tried it with the full path name on the gnatchop
and found it and got the same pragma. So the above steps are really
not complete if you want to use the fully integrated emacs and gnat
(gdb is really the one who has a problem finding the source).
What I would like to see for "librarian" support is switch added to
gnatbl and gcc that would allow me to define how to map a unit name
to a file name. If the option is not used, then use the current
method. It seems this would solve most everyone's complaint about
the naming convention of the source because they could do it their
way instead of living with someone elses standard.
I don't know much about debuggers -- other than how to use them --
but when source is compiled is the complete path and file name of the
source saved in the object code as debugging info? So given the above
wish were to come true, would gcc use the full path and file name of
the source being used in the object file allowing gdb to always find
the source file correctly? Or do you always have to use directory and
path commands in gdb? Either is fine I'm just curious, and it seems
it would be better for the debugger if I always used the directory and
path commands.
Even when I use the gnatchop -r, gdb still looks for the .adb files
instead of the original. Want proof:
······@hades : gnatchop -r -w hello_world.ada
splitting hello_world.ada into:
hello_world.adb
······@hades : more hello_world.adb
pragma Source_Reference (000001, "hello_world.ada");
.
.
.
······@hades : gcc -c -ggdb hello_world.adb
······@hades : gnatbl hello_world.ali
······@hades : rm hello_world.adb
······@hades : gdb hello_world
GDB is free software and you are welcome to distribute copies of it
under certain conditions; type "show copying" to see the conditions.
There is absolutely no warranty for GDB; type "show warranty" for details.
GDB 4.14 (sparc-sun-solaris2.3),
Copyright 1995 Free Software Foundation, Inc...
(gdb) break _ada_hello_world
Breakpoint 1 at 0x1e4cc: file hello_world.adb, line 12.
(gdb) run
Starting program: /home/afniii/Temporary_Space/Ada_Temp/hello_world
Breakpoint 1, _ada_hello_world () at hello_world.adb:12
hello_world.adb:12: No such file or directory.
(gdb)
It still looks for the .adb file -- I deleted .adb so that the error
would be obvious. This is the generic version of gdb NOT the Ada
aware version -- because it supported only with SunOS 4.1.3 (when I
last checked several weeks ago).
From: Robert I. Eachus
Subject: Re: Ethics in programming (was Re: Poetry in programming (Was: Re: Comparing OO against Structured Methodologies))
Date:
Message-ID: <EACHUS.95Oct25185422@spectre.mitre.org>
In article <···············@schonberg> ·····@cs.nyu.edu (Robert Dewar) writes:
> Fair enough, but let's look at this issue in a little calmer
> manner. First, if you want a system with a central library, where
> there is no connection between file names and unit names, this is
> easily achievable with a simple set of shell scripts. Jean Pierre
> Rosen has for instance written such a system as a set of simple
> REXX scripts for OS/2. Of course if you adopt this approach then
> you get all the problems (maybe Jay thinks they are advantages in
> his view) of having to follow a specific order of compilation.
> It would also be quite easy to optionally have a mechanism of
> using a centralized file which did file name translations. Indeed
> this was part of the original design.
> The advantage is that you break the need for a connection between
> the names. The disavdnatgae is that you now have a centralized
> data structure, which existing configuration management systems
> don't know about, and which has to handle common access from
> parallel compilations. Also, if this file is updated by
> compilation, you also have introduced an order of compilation
> requirement.
I think, Robert, that you have indicated a third solution which
may be the preferred one. Don't lose the current mode of operation,
but provide an additional mode where you use a library file which
contains a file to unit mapping BUT NO OTHER INFORMATION. In this
mode you want to be able to say "ada *.spec.a -lib MY_LIB" or
whatever. Translated that should mean "if adding all the files with
names of the form *.spec.a from this directory into the Ada library
MY_LIB produces a consistant library do so. Otherwise provide error
messages and do nothing. This would be a huge help in building large
systems, eliminate (actually replace) the .ali files and not take much
implementation work at all. You just have another way (table-lookup)
for converting between unit names and file names, and code to update
(write) the library after a successful compilation.
The chief advantage of this on large projects is that versioning
is much easier: each version corresponds to a different library file.
--
Robert I. Eachus
with Standard_Disclaimer;
use Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...