From: Robert Elton Maas
Subject: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <7152@public.BTR.COM>
Since I got on Usenet in March (and before when I could kluge NNTP
access via TELNET) I've browsed the *.jobs.offered newsgroups, but most
of the jobs require more C experience than I have, and there are hardly
ever any LISP jobs offered.
So now I thought I'd go at it from the other end by posting to a
lisp-related group asking if anybody can earn a living programming in
LISP. (I have about ten years LISP programming experience, mostly at
Stanford.) If anyone is actually earning a living in industry (not a
university) programming in LISP, could you please tell me how you found
the job and what prior experience was required in addition to being
good at LISP programming?
Or is LISP just a toy language that is wonderful to program in and is
used for A.I. research in universities, but is not suitable for earning
a living in the "real world"?
From: Clinton Hyde
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <CHYDE.92Jun22112756@pecos.ads.com>
I've been "earning a living" doing Lisp exclusively since late 84. i'd
guess that's long enough to meet yr expectations. i suspect many other
readers here have also. if you need a lisp job, I am interviewing, and
I can point to two other employers nearby (Wash DC) who are also
looking for lisp programmers.
-- clint
--
Clint Hyde "Give me a LispM or give me death!" -- jwz
Advanced Decision Systems Internet: ·····@chesapeake.ads.com
2111 Wilson Blvd #800
Arlington, VA 22201 (703) 875-0327
In article <····@public.BTR.COM> ···@btr.btr.com (Robert Elton Maas) writes:
Since I got on Usenet in March (and before when I could kluge NNTP
access via TELNET) I've browsed the *.jobs.offered newsgroups, but most
of the jobs require more C experience than I have, and there are hardly
ever any LISP jobs offered.
*.jobs.offered is a valuable service, but is hardly representative of the
entire computer job market (IMHO).
So now I thought I'd go at it from the other end by posting to a
lisp-related group asking if anybody can earn a living programming in
LISP. (I have about ten years LISP programming experience, mostly at
Stanford.) If anyone is actually earning a living in industry (not a
university) programming in LISP, could you please tell me how you found
the job and what prior experience was required in addition to being
good at LISP programming?
I have been doing Lisp in this group since about '85. We have two open
positions, but the Lab is requiring job searches internally before
advertizing the jobs externally.
Or is LISP just a toy language that is wonderful to program in and is
used for A.I. research in universities, but is not suitable for earning
a living in the "real world"?
We build large simulation models in Lisp that have found a fair amount of
use in the "real world" of both the DOE and with technology transfer to
industry.
My opinion is that there will be a resurgance of interest in Lisp. Lisp got
a bit of tarnish from all the hype of the AI fad of the mid-80s. It also
got a lot of solid technology from the interest of that time that is now
comming to fruition. When the life-cycle costs of a project done in Lisp
are compared to those of a project done in (e.g.) C++, I believe that the
costs of Lisp are far outweighed by the benefits of the language.
You might try to get to LUV-92 in San Diego this August. You WILL find
a lot of examples of Lisp being used for "real work" there.
Skip Egdorf
···@lanl.gov
From: Thomas Pole
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <1992Jun24.182357.24012@software.org>
The discussion of Lisp jobs and Lisp job-hunters is one
I've been thinking a lot about the past few months. Among the
board members of the Association for Lisp Users, we considered
putting together some sort of resource catalog of Lisp vendors,
Lisp users, and Lisp workers.
This was prompted by complaints we had heard that
employers are having a difficult time finding Lisp programmers,
and that Lispers are expensive because they demand high pay.
My own experience has been that I frequently meet Lispers
who are writing 'C' or Ada or even Cobol because they can't
even find a Lisp opening, much less a Lisp offer.
I also meet people using Lisp at work without
any training or background in Lisp. There was a requirement
to use Lisp, or maintain a Lisp system, and a 'C' programmer
was given the job.
Obviously, there is a commmunication gap here.
The answer is not to say "well, the job opening list
is not the whole world of computers." I live in the DC area,
I read the local paper regularly, I read the job postings regularly,
and I had no idea that there were three firms in the DC area hiring
Lisp programmers. I found out in one of the postings in this discussion.
I agree that Lisp is ready, and able to make a commercial comeback.
I think it can become one of the mainstream development languages of the 90's.
I also believe that it should be one of the mainstream languages of the 90's.
I use CL almost exclusively in my work here at SPC, and the companies
I work with have no qualms about working with Lisp-based tools. (managements
contrary feelings not-withstanding) But, if we don't communicate among
ourselves, we will _not_ get the message out about the commercial
vialbility of CL to anyone else.
I open the floor to all of you, how do we:
(1) Make Lisp opportunities (jobs, classses, whatever)
public knowledge.
(2) Ensure that we have a next generation of Lisp programmers coming
out of the Universities.
(3) Make Lisp development environements (full CL, not 85% Lisp's)
available on common platforms (80x86/MSDOS especially), at
prices competitive with C++ and others.
From: Bob Riemenschneider
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <RAR.92Jun24143931@birch.csl.sri.com>
In article <······················@software.org> ····@software.org (Thomas Pole) writes:
I open the floor to all of you ...
Before this discussion gets going, would you post the charter of the
Association of Lisp Users? Is the organization committed to addressing
any of these issues?
-- rar
From: Robert Maas, see http://tinyurl.com/uh3t
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <REM-2005sep03-007@Yahoo.Com>
> http://groups.google.com/group/comp.lang.lisp/msg/f58e436954aa934a
> From: ····@software.org (Thomas Pole)
> The discussion of Lisp jobs and Lisp job-hunters is one
> I've been thinking a lot about the past few months. Among the
> board members of the Association for Lisp Users, we considered
> putting together some sort of resource catalog of Lisp vendors,
> Lisp users, and Lisp workers.
Did you, or anyone, ever get around to anything like that? If so, where
is it located online?
> This was prompted by complaints we had heard that
> employers are having a difficult time finding Lisp programmers,
> and that Lispers are expensive because they demand high pay.
Lispers are not expensive at all, and that mis-perception by potential
employers is a serious problem if you are correct in reporting such
perception. Do you, or does anyone reading this thread, have any
specific figure of how much $/hr employers believe Lispers require?
And how much lower would Lispers need to go before they'd match what
employers are willing to pay?
> I agree that Lisp is ready, and able to make a commercial comeback.
> I think it can become one of the mainstream development languages of the 90's.
> I also believe that it should be one of the mainstream languages of the 90's.
Obviously your prediction was wrong. Even now in 2005, there haven't
been any advertised Lisp programming jobs in more than ten years, while
there have been many advertised programming jobs using the mainstream
languages (Java, C++, Visual Basic .NET, etc.).
For example, the most recent job ad in this area that even mentionned
Lisp was more than a year ago:
http://groups.google.com/group/ba.jobs.offered/msg/a9e72b113264bc18
Message-ID: <···················@newssvr27.news.prodigy.com>
and it wasn't even a programming job, it was a job for "QA Engineer",
which required specific experience with Windows 2000, Linux, Solaris,
MSSQL, Oracle, QuickTest Pro, Python, VBScript, Perl, SQL, XML, ETL,
JSP, HTML, and then way down after all that it merely asks for
"Knowledge of Lisp/Scheme Oracle, SQL Server, PostgreSQL, MySQL, DB2",
no actual Lisp experience required or given any sort of priority.
Before that, the previous job that mentionned Lisp
http://groups.google.com/group/ba.jobs.offered/msg/d23d7a2cb717d3ee
Message-ID: <····························@posting.google.com>
required: expertise in C++, a (potentially) strong interest in human
simulation and human factors modeling, and then finally only helpful,
not required: experience with Lisp, software configuration management
(version control, builds), and product development.
Before that, the previous job that mentionned Lisp was all the way back
in 2000:
http://groups.google.com/group/ba.jobs.offered/msg/76d2f6a2a6fe7523
Message-ID: <··················@client>#1/1
(SmallTalk/C++ java)
The last time a job ad mentionned Lisp in the Subject field was in
1998:
http://groups.google.com/group/ba.jobs.offered/msg/4c3661596cc04a91
Message-ID: <···············@news.ISWEST.COM>#1/1
But the primary requirement there is "A minimum of two to three years
of experience in creating and shipping commercial software." which
excludes pretty much all Lispers.
The most recent job specifically for Lisp was also in 1998:
http://groups.google.com/group/ba.jobs.offered/msg/b47980af9d39ea66
Message-ID: <···············@news.micron.net>#1/1
but that job seems to be mostly implementing lisp, not using lisp. The
only thing like that I ever did was to write the save/restore
environment code for PSL (Portable Standard Lisp), which had nothing
more to do than call the garbage collector to compact the heap, then
simply write out all the data areas such as heap to disk file,
accompanied by header info to tell how large each block was, and
likewise read such a file back into PSL at start-up to restore the old
state again.
The most recent actual Lisp application-programming job seems to be this:
CA-Mtn.View-AI / LISP PROGRAMMER / COGNITIVE MODELER / NASA ...
Duties will largely consist of analysis and programming in common
LISP, along with
CLIM programming and the analysis and evaluation of data using
statistical ...
comp.jobs.offered - Aug 11 1998, 9:08 am by Sue Cox
but Google Groups Broken Beta can't find it:
Topic not found.
We're sorry, but we were unable to find the topic you were looking for.
Perhaps the URL you clicked on is out of date or broken?
I've scanned all the way back to the beginning of 1997 and don't see a
single job ad for regular Lisp programming (not Emacs nor CAD nor impl.).
So that covers 80% of the ten years I claim not a single Lisp
programming job, because there haven't been any such ads, not because
there were ads I missed.
> I open the floor to all of you, how do we:
> (1) Make Lisp opportunities (jobs, classses, whatever)
> public knowledge.
If only there were such opportunities, sigh.
> (2) Ensure that we have a next generation of Lisp programmers coming
> out of the Universities.
No!! We already have too many unemployed Lispers, don't need any more!
Find Lisp programming jobs for all us unemployed Lispers first!
In article <·················@Yahoo.Com>,
·······@Yahoo.Com (Robert Maas, see http://tinyurl.com/uh3t) wrote:
> Even now in 2005, there haven't
> been any advertised Lisp programming jobs in more than ten years
That is not true.
It does seem to be true that in the last ten years there haven't been
any Lisp programming jobs listed in ba.jobs.offered, which seems to be
the only place you've looked. There also don't seem to be any Lisp
programming jobs that list sniveling, whining, or feeling sorry for
onesself as desired qualifications. (Probably just as well. No doubt
you'd be rejected as overqualified.)
There also haven't been any openings for blacksmiths, underwater basket
weavers, carriage riggers, nose-pickers, or pina colada tasters either.
Stupid gits. I happen to be a fucking brilliant pina colada taster, but
I can't find a job tasting pina coladas no matter what I do. I wonder
why that is? Sometimes life just seems so damned unfair.
Say, have you considered looking for work as a longshoreman? Carrying
that two-ton parenthesis-shaped chip on your shoulder for as long as you
have should have built up some serious upper-body strength by now.
rg
Ron Garret <·········@flownet.com> writes:
> Say, have you considered looking for work as a longshoreman? Carrying
> that two-ton parenthesis-shaped chip on your shoulder for as long as you
> have should have built up some serious upper-body strength by now.
Ha ha ha ... Lisp jobs are not too hard so come by ... work as a
longshoreman on the other hand *is* hard to get (in California, at
least). I found three Lisp jobs in the time it took one of my friends
to get a callback for a job as a longshore B-man.
--
/|_ .-----------------------.
,' .\ / | Free Mumia Abu-Jamal! |
,--' _,' | Abolish the racist |
/ / | death penalty! |
( -. | `-----------------------'
| ) |
(`-. '--.)
`. )----'
Maybe there's a better way to go about it. Look for a job you want to do,
a job you can do, then use Lisp to do it. I've always programmed in pretty
much whatever language I wanted - and now that I'm learning Lisp I'm
programming 75% of my work in Lisp. I didn't ask if I could, I just do
it. As long as I'm producing what they want they don't really care. And
what I'm finding about Lisp is that even though I'm way down on the
learning curve, I can tweak and rearrange my codes much more quickly than
in the past with C and Perl, so I'm getting more done (or another way of
looking at it is I have more free time to screw around learning more Lisp).
-Jeff Cunningham
From: Robert Maas, see http://tinyurl.com/uh3t
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <REM-2005sep05-002@Yahoo.Com>
> From: Jeff Cunningham <·······@cunningham.net>
> Look for a job you want to do, a job you can do, then use Lisp to do
> it.
I'm not aware of that being an option. Virtually all job ads not only
specify the kind of application, but also specify the programming
language, which is *not* Lisp in 99.999% of cases. Applying for such a
job, with the intention of writing the program in Lisp instead of the
language they specified in the ad, seems a stupid idea, where I'd get
fired the very first day when they glanced at my work and saw Lisp
s-expressions instead of whatever language they wanted me to use.
> now that I'm learning Lisp I'm programming 75% of my work in Lisp.
Is anyone willing to pay you for that Lisp software??
> As long as I'm producing what they want they don't really care.
Is this in-house-use utility/research software, or server-side
software, or something actually installed on customer machines which
requires each customer purchase a licensed copy of the Lisp system
before they can run your software?
From: Christopher C. Stacy
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <uk6hvpb5q.fsf@news.dtpq.com>
·······@Yahoo.Com (Robert Maas, see http://tinyurl.com/uh3t) writes:
> > From: Jeff Cunningham <·······@cunningham.net>
> >
> > now that I'm learning Lisp I'm programming 75% of my work in Lisp.
>
> Is this in-house-use utility/research software, or server-side
> software, or something actually installed on customer machines which
> requires each customer purchase a licensed copy of the Lisp system
> before they can run your software?
You are arguing a false dilemma; not all software (in particular,
not all commercial Lisp systems) have runtime redistribution fees.
I don't know what Jeff is using, but I distribute executable Lisp
programs to end users (on a variety of platforms). Since you read
this newsgroup, you should know that this is common, since other
people occasionally post notices of free downloadable programs
(usually for MS Windows) that are written in Lisp.
From: Stefan Scholl
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <dfm7o5$9e1$00$1@news.t-online.com>
Christopher C. Stacy <······@news.dtpq.com> wrote:
> programs to end users (on a variety of platforms). Since you read
> this newsgroup, you should know that this is common, since other
> people occasionally post notices of free downloadable programs
> (usually for MS Windows) that are written in Lisp.
Can't remember such a post. I just use one Lisp programm for Windows:
The Regex Coach
·······@Yahoo.Com (Robert Maas, see http://tinyurl.com/uh3t) writes:
>> From: Jeff Cunningham <·······@cunningham.net>
>> Look for a job you want to do, a job you can do, then use Lisp to do
>> it.
>
> I'm not aware of that being an option. Virtually all job ads not only
> specify the kind of application, but also specify the programming
> language, which is *not* Lisp in 99.999% of cases. Applying for such a
> job, with the intention of writing the program in Lisp instead of the
> language they specified in the ad, seems a stupid idea, where I'd get
> fired the very first day when they glanced at my work and saw Lisp
> s-expressions instead of whatever language they wanted me to use.
Of course, you have to be smart about it.
Like for example Linj:
http://www.evaluator.pt/downloads/tutorial.html
Imagine the productivity you could display modifying without telling a
couple of lisp macros, and regenerating megaLOCs of updated Java code!
--
__Pascal Bourguignon__ http://www.informatimago.com/
Until real software engineering is developed, the next best practice
is to develop with a dynamic system that has extreme late binding in
all aspects. The first system to really do this in an important way
is Lisp. -- Alan Kay
As a happy Linj user, I second that. Or just the fact we often hear on
this group, that it's easier to tweak Lisp code--add a middle loop
around an inner loop, make a function return a second value that need
to use sometimes but not always. Then the changes are all compiled out
to Java for you.
From: Robert Maas, see http://tinyurl.com/uh3t
Subject: Linj (was: Any LISP programmers able to earn a living in the ...)
Date:
Message-ID: <REM-2005sep14-003@Yahoo.Com>
> From: Pascal Bourguignon <····@mouse-potato.com>
> > ... Applying for such a
> > job, with the intention of writing the program in Lisp instead of the
> > language they specified in the ad, seems a stupid idea, where I'd get
> > fired the very first day when they glanced at my work and saw Lisp
> > s-expressions instead of whatever language they wanted me to use.
> Of course, you have to be smart about it.
> Like for example Linj:
> http://www.evaluator.pt/downloads/tutorial.html
> Imagine the productivity you could display modifying without telling a
> couple of lisp macros, and regenerating megaLOCs of updated Java code!
The extra productivity in Lisp, compared to Java, is not in the syntax
so much as in the extra capability:
- read/print: equivalence between internal data structures and syntax.
- eval: ability to generate code expressions at runtime and execute them.
- runtime symbols with value/function/properties, auto-interned when read.
- keyword parameters and multiple return values fully integrated.
Transforming syntax between Java and Lisp doesn't provide any of these
Lisp-only capabilities. For capabilities already in Java, there doesn't
seem much point in converting to/from Lisp syntax. So I don't see how
Linj would support greatly increased productivity compared to just
coding in Java. GNU Emacs supports Java syntax almost as well as it
supports Lisp syntax, so ease of editing isn't a major reason for
preferring Lisp syntax when coding for Java compiler.
I didn't mention anonymous inner methods (in Java) or lexical closures
(in Lisp), because I don't know how the two are related modulo Linj.
Can somebody write arbitrary MAPCAR etc. expressions, or LOOP
expressions, with references to non-lambda lexical variables in the
containing function, in Lisp, and have them converted to appropriate
correct Java code? If so, that might be a possible argument for using
Linj, since coding that sort of task directly in Java is somewhat
tricky. On the other hand, writing such code is totally natural in
Lisp, done "all the time", and if Linj can't convert it to valid Java
that would be a solid argument against ever using Linj. So does any
experience Linj user know how well Linj converts such mapping/LOOP
expressions to Java?
But the question of programming in Lisp instead of Java is rather moot,
since I can get the job done (in most cases) almost as well in Java as
in Lisp. The problem would be if I was hired to write in C or C++ etc.,
and found it virtually impossible to get any useful productivity
without using Lisp (or Java), so I went ahead and wrote Lisp instead of
C/C++ code, where Linj wouldn't be able to save me from getting fired.
It'd be a really big gamble hoping my boss would notice my greatly
increased productivity over the C/C++ programmers, and decide to
forgive me for not following instructions and for lying on my job
application etc.
Robert Maas, see http://tinyurl.com/uh3t wrote:
> The extra productivity in Lisp, compared to Java, is not in the syntax
> so much as in the extra capability:
> - read/print: equivalence between internal data structures and syntax.
You think translation couldn't do this?
> - eval: ability to generate code expressions at runtime and execute them.
Hm, rare, I'd say.
> - runtime symbols with value/function/properties, auto-interned when read.
> - keyword parameters and multiple return values fully integrated.
Sure, but translation can handle this transparently, if slowly.
> Transforming syntax between Java and Lisp doesn't provide any of these
> Lisp-only capabilities. For capabilities already in Java, there doesn't
> seem much point in converting to/from Lisp syntax. So I don't see how
> Linj would support greatly increased productivity compared to just
> coding in Java. GNU Emacs supports Java syntax almost as well as it
> supports Lisp syntax, so ease of editing isn't a major reason for
> preferring Lisp syntax when coding for Java compiler.
Lisp syntax, JVM output. Also, you can have macros that write loads of
Java for you. Oh, and more readable code (better indentation, fewer
verbose distractions) and better editing facilities at the sexp level.
> I didn't mention anonymous inner methods (in Java) or lexical closures
> (in Lisp), because I don't know how the two are related modulo Linj.
> Can somebody write arbitrary MAPCAR etc. expressions, or LOOP
> expressions, with references to non-lambda lexical variables in the
> containing function, in Lisp, and have them converted to appropriate
> correct Java code? If so, that might be a possible argument for using
I suspect they could be transformed into for(;;) loops.
> Linj, since coding that sort of task directly in Java is somewhat
> tricky. On the other hand, writing such code is totally natural in
> Lisp, done "all the time", and if Linj can't convert it to valid Java
> that would be a solid argument against ever using Linj. So does any
> experience Linj user know how well Linj converts such mapping/LOOP
> expressions to Java?
(Haven't tried it yet.)
> But the question of programming in Lisp instead of Java is rather moot,
> since I can get the job done (in most cases) almost as well in Java as
> in Lisp. The problem would be if I was hired to write in C or C++ etc.,
Lucky you. I don't mind Java, but Lisp is so much more convenient.
> and found it virtually impossible to get any useful productivity
> without using Lisp (or Java), so I went ahead and wrote Lisp instead of
> C/C++ code, where Linj wouldn't be able to save me from getting fired.
> It'd be a really big gamble hoping my boss would notice my greatly
> increased productivity over the C/C++ programmers, and decide to
> forgive me for not following instructions and for lying on my job
> application etc.
Heh.
--
My mouth says the words, my brain is thinking monstertrucks.
Joey (Friends)
From: Kent M Pitman
Subject: Re: Linj (was: Any LISP programmers able to earn a living in the ...)
Date:
Message-ID: <uy85ye1rb.fsf@nhplace.com>
·······@Yahoo.Com (Robert Maas, see http://tinyurl.com/uh3t) writes:
> The extra productivity in Lisp, compared to Java, is not in the syntax
> so much as in the extra capability:
> - read/print: equivalence between internal data structures and syntax.
> - eval: ability to generate code expressions at runtime and execute them.
> - runtime symbols with value/function/properties, auto-interned when read.
> - keyword parameters and multiple return values fully integrated.
> Transforming syntax between Java and Lisp doesn't provide any of these
> Lisp-only capabilities.
Even within the limited scope of virtues you've listed here...
- Java doesn't have macros. That's value already. Java doesn't have
macros (or didn't last I looked).
- Beyond just having macros, the ability to transform macros and see
them expanded and pretty-printed interactively seems additionally
worthwhile. In C/C++ you "have macros", but it's hard to see what
they will expand to.
Robert Maas, see http://tinyurl.com/uh3t escreveu:
> > From: Pascal Bourguignon <····@mouse-potato.com>
> > > ... Applying for such a
> > > job, with the intention of writing the program in Lisp instead of the
> > > language they specified in the ad, seems a stupid idea, where I'd get
> > > fired the very first day when they glanced at my work and saw Lisp
> > > s-expressions instead of whatever language they wanted me to use.
> > Of course, you have to be smart about it.
> > Like for example Linj:
> > http://www.evaluator.pt/downloads/tutorial.html
> > Imagine the productivity you could display modifying without telling a
> > couple of lisp macros, and regenerating megaLOCs of updated Java code!
>
> The extra productivity in Lisp, compared to Java, is not in the syntax
> so much as in the extra capability:
> - read/print: equivalence between internal data structures and syntax.
Maybe I don't understand what you are saying but, as you surely know,
that equivalence doesn't work on several Common Lisp datatypes,
including hashtables, functions, packages, and, more importantly,
instances of CLOS classes. Most of my programs use (among others) a
combination of the previous datatypes so I cannot count on that
"feature" and if I want it, I have to implement it myself. Just like
in Java.
Linj, OTHO, attempts to implement read/print even for instances of
classes: you just have to include :make-parse-method as a class option
and you will get consistent read and print methods for the class'
instances. For example, consider the following definition:
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(defclass complex ()
((x :initarg :x :reader real-part)
(y :initarg :y :reader imaginary-part))
(:make-parse-method t))
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
When you translate this to Java, you get:
//////////////////////////////////////////////////////////////////////
class Complex extends Object {
// constructors
public Complex(Object x, Object y) {
this.x = x;
this.y = y;
}
// accessors
public Object realPart() {
return x;
}
public Object imaginaryPart() {
return y;
}
// methods
public String toString() {
StringBuffer buf = new StringBuffer();
buf.append("#S(Complex :x ");
buf.append(x);
buf.append(" :y ");
buf.append(y);
buf.append(")");
return buf.toString();
}
public static Complex parse(Cons internalRawData) throws
ClassParseException {
Cons internalRawDataKeys = internalRawData.rest();
if (Symbol.intern("Complex") != internalRawData.first()) {
throw new ClassParseException("Error parsing class " +
Symbol.intern("Complex") +
" object: " +
internalRawData);
} else {
return new
Complex(internalRawDataKeys.getf(Symbol.intern(":x"), null),
internalRawDataKeys.getf(Symbol.intern(":y"), null));
}
}
// slots
protected Object x;
protected Object y;
}
//////////////////////////////////////////////////////////////////////
As you can see, you get the ability to print and, even better, to read
instances of the complex class. There's another option that, for the
moment, only works in the 'print' direction: make-xml-parse-method.
You can guess what it does.
The 'read' part of the process is partly implemented by the
linj-reader, a class (implemented in Linj, obviously) that attempts to
read several Linj data-types according to Common Lisp rules. It
includes lists, symbols, numbers (including fractions) and, calls to
the above 'parse' method. It also handles (a generalized version of)
#1= and #1# so that you can read circular structures.
> - eval: ability to generate code expressions at runtime and execute them.
I wonder how many Common Lisp programmers use eval on a regular basis.
I don't, so Linj doesn't implement it. But nothing prevents you from
adding one to Linj. Will it make sense in the Java side? I doubt.
> - runtime symbols with value/function/properties, auto-interned when
> read.
Linj implements symbols but they don't have function cell. But they
do have value and plist cell. And they are auto-interned by the Linj
reader.
> - keyword parameters and multiple return values fully integrated.
Linj has &keyword parameters (and &optional and &rest and &body and
init-forms and supplied-p) although there are a few rare cases that
have different semantics from Common Lisp.
Multiple return values are also a Linj feature. Not as flexible as in
Common Lisp, but usable nevertheless.
So, it seems to me that Linj implements, at least, part of the "extra
productivity" features list that you mention. However, I find it
strange that your list doesn't include macros: IMHO, they are one of
the biggest productivity features of Lisp. And Linj has them, of
course, and the good news is that they are implemented by the Common
Lisp side so you have all the power of Common Lisp at your disposal
while macro-expanding your Linj macros. Believe me: that's where you
will get a big Java productivity boost.
> Transforming syntax between Java and Lisp doesn't provide any of these
> Lisp-only capabilities.
It depends on the kind of "syntax transformation" that you use.
> For capabilities already in Java, there doesn't seem much point in
> converting to/from Lisp syntax. So I don't see how Linj would
> support greatly increased productivity compared to just coding in
> Java.
Java doesn't have macros or keyword parameters or any of the good Linj
features that I value. Maybe you value different things and Linj
isn't the right tool for you. But it sure is for me and it allowed me
to stop writing Java code a long time ago.
> GNU Emacs supports Java syntax almost as well as it supports Lisp
> syntax, so ease of editing isn't a major reason for preferring Lisp
> syntax when coding for Java compiler.
No one mentioned "ease of editing" as a major reason. It isn't for me
anyway, although, in Emacs, I do find it easier to edit Linj than
Java. But note that there are lots of good things that you can find on
a good
Java editor (such as Eclipse) that you will not have while editing
Linj code in Emacs. Eclipse simplifies some of the tedious tasks of
Java programming.
> I didn't mention anonymous inner methods (in Java) or lexical closures
> (in Lisp), because I don't know how the two are related modulo Linj.
They are related.
Here is the cannonical example in Linj:
(defun adder (x)
(function (lambda (y) (+ x y))))
(defun main ()
(let ((add3 (adder 3)))
(print (funcall add3 5))))
and here is the translation:
public static Function adder(final Bignum x) {
return new Function() {
public Object funcall(Object genericY) {
Bignum y = (Bignum)genericY;
return x.add(y);
}};
}
public static void main(String[] outsideArgs) {
Function add3 = adder(Bignum.valueOf(3));
System.out.print("" + '\n' + add3.funcall(Bignum.valueOf(5)) +
" ");
}
You'll notice the presence of some supporting classes such as Bignum
and Function. Here's another example where you map a procedure (that
is, a function that doesn't return values) over an abstract list (a
Java class):
(defun map-list (l f)
(declare (abstract-list l) (procedure f))
(dotimes (i (size l))
(funcall f (get l i))))
Let's use it:
(defun print-list (l)
(declare (abstract-list l))
(map-list l #'(lambda (e) (format t "Elem: ~A~%" e))))
(defun copy-vector (v)
(declare (Vector v))
(let ((copy (new 'Vector)))
(map-list v #'(lambda (e) (add-element copy e)))
copy))
and here are the translations:
public static void mapList(AbstractList l, Procedure f) {
int limit = l.size();
for (int i = 0; i < limit; ++i) {
f.funcall(l.get(i));
}
}
public static void printList(AbstractList l) {
mapList(l,
new Procedure() {
public void funcall(Object e) {
System.out.print("Elem: ");
System.out.println(e);
}});
}
public static Vector copyVector(Vector v) {
final Vector copy = new Vector();
mapList(v,
new Procedure() {
public void funcall(Object e) {
copy.addElement(e);
}});
return copy;
}
Notice, in the above example, that the 'copy' variable had to be
(automatically) declared 'final' because it was captured by the
closure.
>
> Can somebody write arbitrary MAPCAR etc. expressions, or LOOP
> expressions, with references to non-lambda lexical variables in the
> containing function, in Lisp, and have them converted to appropriate
> correct Java code?
'Arbitrary' expressions, no. But a large number of situations are
more or less convertable to Java. There are limitations, though. The
most obvious is that, inside a closure, you can't assign to captured
variables. Other, less obvious limitation, is that you might need to
add a type declaration here and there.
However, I must stress that altough you can program using closures,
that's not what the careful Linj programmer will do, in general. The
Linj programmer wants to give their clients the feeling that the
generated Java code was produced by a (highly productive) Java
programmer. With this goal in his mind, the Linj programmer will
develop a style where he naturaly avoids Lisp forms that translate
into hard-to-understand Java forms.
(Re)using the above example, the Linj programmer might use a macro in
place of a higher-order function (for simplicity, let's forget name
capture, repeated evaluations, etc):
(defmacro for-list ((var l) &body body)
`(dotimes (i (size ,l))
(let ((,var (get ,l i)))
,@body)))
Now, let's rewrite the examples:
(defun print-list2 (l)
(declare (abstract-list l))
(for-list (e l)
(format t "Elem: ~A~%" e)))
(defun copy-vector2 (v)
(declare (Vector v))
(let ((copy (new 'Vector)))
(for-list (e v)
(add-element copy e))
copy))
and let's see the translation
public static void printList2(AbstractList l) {
int limit = l.size();
for (int i = 0; i < limit; ++i) {
Object e = l.get(i);
System.out.print("Elem: ");
System.out.println(e);
}
}
public static Vector copyVector2(Vector v) {
Vector copy = new Vector();
int limit = v.size();
for (int i = 0; i < limit; ++i) {
Object e = v.get(i);
copy.addElement(e);
}
return copy;
}
As you can see, it looks more similar to what a normal Java programmer
would write.
BTW, I didn't test any of the above. There might be errors.
> If so, that might be a possible argument for using Linj, since
> coding that sort of task directly in Java is somewhat tricky. On the
> other hand, writing such code is totally natural in Lisp, done "all
> the time", and if Linj can't convert it to valid Java that would be
> a solid argument against ever using Linj. So does any experience
> Linj user know how well Linj converts such mapping/LOOP expressions
> to Java?
I guess I'm an experienced Linj user. All I can say is that it works
for me. I don't claim that it will work for you or for anyone else.
And if you program with lots and lots of higher-order function, then
you'll find some problems and, moreover, the result of the translation
will surely look really strange for the normal Java programmer.
> But the question of programming in Lisp instead of Java is rather moot,
> since I can get the job done (in most cases) almost as well in Java as
> in Lisp.
That's not my experience. Besides, I like to have some fun in what I
do. Java is annoying. Common Lisp isn't. That makes a huge
difference in my productivity.
There are tasks, however, that are much easier to do in Java than in
Common Lisp but that's only because of the very large Java API and the
incredibly large amount of libraries that are available for Java and
not for Common Lisp.
> The problem would be if I was hired to write in C or C++ etc., and
> found it virtually impossible to get any useful productivity without
> using Lisp (or Java), so I went ahead and wrote Lisp instead of
> C/C++ code, where Linj wouldn't be able to save me from getting
> fired. It'd be a really big gamble hoping my boss would notice my
> greatly increased productivity over the C/C++ programmers, and
> decide to forgive me for not following instructions and for lying on
> my job application etc.
I don't have a solution to that problem. But if I'm forced to write
C/C++ I'll surely try to build something similar to Linj that, at the
very least, provides real macro-capabilities to the language.
António Leitão.
··············@evaluator.pt writes:
> It also handles (a generalized version of) #1= and #1# so that you
> can read circular structures.
Generalized how? (I didn't find this in the manual, but perhaps I
didn't look close enough.)
Lars Brinkhoff escreveu:
> ··············@evaluator.pt writes:
> > It also handles (a generalized version of) #1= and #1# so that you
> > can read circular structures.
>
> Generalized how? (I didn't find this in the manual, but perhaps I
> didn't look close enough.)
It's 'generalized' simply to allow #<name>= and #<name># where name can
be any token and not only an unsigned decimal integer as in Common
Lisp. This allows me to write using a kind of 'let'-style that makes
the shared parts more obvious (at least, to me), e.g.
(#triple=(#pair=(1 2) #pair# #pair#) #triple#) =>
(((1 2) (1 2) (1 2)) ((1 2) (1 2) (1 2)))
This also works while reading instances of classes.
Meanwhile, after some testing (and looking at the code), I found that
my implementation is incomplete. In fact, at the moment, it cannot
handle circular structures. I'm sorry for spreading misinformation.
It only handles shared (non-circular) structures.
Just like many other things in Linj, the sharp equal macro-character
was implemented just to solve one particular problem that I had and I
didn't even mentioned it in the manual. Unfortunately I did not
provide a complete solution at the time (I didn't need it). Maybe one
day I'll take a look at it. Anyway, if anyone wants to implement it,
he/she just has to change the linj-reader file (which is written in
Linj).
However, I must stress that reading Java objects with Lisp syntax is
funny but not something that fits well in the Java world. That's why I
also made a very small attempt at providing the same capabilities but
using the horrible XML syntax. The print part was easy but the reader
would require a XML parser and I decided to stop there.
Antonio Leitao
> I'm not aware of that being an option. Virtually all job ads not only
> specify the kind of application, but also specify the programming
> language, which is *not* Lisp in 99.999% of cases. Applying for such a
> job, with the intention of writing the program in Lisp instead of the
> language they specified in the ad, seems a stupid idea, where I'd get
> fired the very first day when they glanced at my work and saw Lisp
> s-expressions instead of whatever language they wanted me to use.
Well, if you look for a job programming Java, sure, I agree with you.
Look for a job doing what they want you programming Java for instead.
;-)
> Is anyone willing to pay you for that Lisp software??
I'm paid very well, thank you.
> Is this in-house-use utility/research software, or server-side
> software, or something actually installed on customer machines which
> requires each customer purchase a licensed copy of the Lisp system
> before they can run your software?
My company doesn't sell software, they sell the products the software
helps produce. Most of the code I write is research related, but not
all of it. If you want to write Lisp, find something that needs to be
done, and do it in Lisp. No one questions successful outcomes unless
they're idiots.
-jeff
On Sat, 03 Sep 2005 17:24:57 -0700, Robert Maas, see
http://tinyurl.com/uh3t wrote:
> The last time a job ad mentionned Lisp in the Subject field was in 1998:
> http://groups.google.com/group/ba.jobs.offered/msg/4c3661596cc04a91
> Message-ID: <···············@news.ISWEST.COM>#1/1
> But the primary requirement there is "A minimum of two to three years of
> experience in creating and shipping commercial software." which excludes
> pretty much all Lispers.
And why would it do that?
On Sat, 03 Sep 2005 17:24:57 -0700, Robert Maas, see
http://tinyurl.com/uh3t wrote:
>> http://groups.google.com/group/comp.lang.lisp/msg/f58e436954aa934a
>> From: ····@software.org (Thomas Pole)
>> The discussion of Lisp jobs and Lisp job-hunters is one
>> I've been thinking a lot about the past few months. Among the
>> board members of the Association for Lisp Users, we considered
>> putting together some sort of resource catalog of Lisp vendors,
>> Lisp users, and Lisp workers.
>
> Did you, or anyone, ever get around to anything like that? If so, where
> is it located online?
>
>> This was prompted by complaints we had heard that
>> employers are having a difficult time finding Lisp programmers,
>> and that Lispers are expensive because they demand high pay.
>
> Lispers are not expensive at all, and that mis-perception by potential
> employers is a serious problem if you are correct in reporting such
> perception. Do you, or does anyone reading this thread, have any
> specific figure of how much $/hr employers believe Lispers require?
> And how much lower would Lispers need to go before they'd match what
> employers are willing to pay?
>
>> I agree that Lisp is ready, and able to make a commercial comeback.
>> I think it can become one of the mainstream development languages of the 90's.
>> I also believe that it should be one of the mainstream languages of the 90's.
>
> Obviously your prediction was wrong. Even now in 2005, there haven't
> been any advertised Lisp programming jobs in more than ten years, while
> there have been many advertised programming jobs using the mainstream
> languages (Java, C++, Visual Basic .NET, etc.).
>
> For example, the most recent job ad in this area that even mentionned
> Lisp was more than a year ago:
> http://groups.google.com/group/ba.jobs.offered/msg/a9e72b113264bc18
> Message-ID: <···················@newssvr27.news.prodigy.com>
> and it wasn't even a programming job, it was a job for "QA Engineer",
> which required specific experience with Windows 2000, Linux, Solaris,
> MSSQL, Oracle, QuickTest Pro, Python, VBScript, Perl, SQL, XML, ETL,
> JSP, HTML, and then way down after all that it merely asks for
> "Knowledge of Lisp/Scheme Oracle, SQL Server, PostgreSQL, MySQL, DB2",
> no actual Lisp experience required or given any sort of priority.
>
> Before that, the previous job that mentionned Lisp
> http://groups.google.com/group/ba.jobs.offered/msg/d23d7a2cb717d3ee
> Message-ID: <····························@posting.google.com>
> required: expertise in C++, a (potentially) strong interest in human
> simulation and human factors modeling, and then finally only helpful,
> not required: experience with Lisp, software configuration management
> (version control, builds), and product development.
>
> Before that, the previous job that mentionned Lisp was all the way back
> in 2000:
> http://groups.google.com/group/ba.jobs.offered/msg/76d2f6a2a6fe7523
> Message-ID: <··················@client>#1/1
> (SmallTalk/C++ java)
>
> The last time a job ad mentionned Lisp in the Subject field was in
> 1998:
> http://groups.google.com/group/ba.jobs.offered/msg/4c3661596cc04a91
> Message-ID: <···············@news.ISWEST.COM>#1/1
> But the primary requirement there is "A minimum of two to three years
> of experience in creating and shipping commercial software." which
> excludes pretty much all Lispers.
>
> The most recent job specifically for Lisp was also in 1998:
> http://groups.google.com/group/ba.jobs.offered/msg/b47980af9d39ea66
> Message-ID: <···············@news.micron.net>#1/1
> but that job seems to be mostly implementing lisp, not using lisp. The
> only thing like that I ever did was to write the save/restore
> environment code for PSL (Portable Standard Lisp), which had nothing
> more to do than call the garbage collector to compact the heap, then
> simply write out all the data areas such as heap to disk file,
> accompanied by header info to tell how large each block was, and
> likewise read such a file back into PSL at start-up to restore the old
> state again.
>
> The most recent actual Lisp application-programming job seems to be this:
> CA-Mtn.View-AI / LISP PROGRAMMER / COGNITIVE MODELER / NASA ...
> Duties will largely consist of analysis and programming in common
> LISP, along with
> CLIM programming and the analysis and evaluation of data using
> statistical ...
> comp.jobs.offered - Aug 11 1998, 9:08 am by Sue Cox
> but Google Groups Broken Beta can't find it:
> Topic not found.
> We're sorry, but we were unable to find the topic you were looking for.
> Perhaps the URL you clicked on is out of date or broken?
>
> I've scanned all the way back to the beginning of 1997 and don't see a
> single job ad for regular Lisp programming (not Emacs nor CAD nor impl.).
> So that covers 80% of the ten years I claim not a single Lisp
> programming job, because there haven't been any such ads, not because
> there were ads I missed.
>
>> I open the floor to all of you, how do we:
>> (1) Make Lisp opportunities (jobs, classses, whatever)
>> public knowledge.
>
> If only there were such opportunities, sigh.
>
>> (2) Ensure that we have a next generation of Lisp programmers coming
>> out of the Universities.
>
> No!! We already have too many unemployed Lispers, don't need any more!
> Find Lisp programming jobs for all us unemployed Lispers first!
Here's one:
http://netfonds.no/stilling.php
--
mvh,
Lars Rune Nøstdal
http://lars.nostdal.org/
·······@Yahoo.Com (Robert Maas, see http://tinyurl.com/uh3t) writes:
> Even now in 2005, there haven't been any advertised Lisp programming
> jobs in more than ten years
One (of several found just fooling around) counter example:
http://seeker.dice.com/jobsearch/servlet/JobSearch?op=101&···············································@activejobs0&c=1&source=1
/Jon
--
'j' - a n t h o n y at romeo/charley/november com
From: Robert Maas, see http://tinyurl.com/uh3t
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <REM-2005sep05-001@Yahoo.Com>
> From: jayessay <······@foo.com>
> http://seeker.dice.com/jobsearch/servlet/JobSearch?op=101&···············································@activejobs0&c=1&source=1
My experience with DICE is that none of the jobs they advertise are
legitimate in the sense of anyone actually responding to resumes sent.
But I went ahead and tried to apply anyway:
- They actively discriminate against anyone without the money to
purchase MicroSoft Word.
- The only space for sending my resume is a single-line textfield, so
the most I could put there was one URL pointing to my resume.
- When I clicked the submit button, I got an error message something
like "can't open file for upload". I've never seen that particular
error message from any other WebSite, so the Dice site seems to be
broken in a unique way.
Now if you can find a recent ad for a Lisp job which doesn't
discriminate against people who don't already have jobs with lots of
money to purcharse MicroSoft products, and where there is actually a
working submit-resume process, I'd be glad to try again.
But thanks for trying anyway.
From: Ulrich Hobelmann
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <3o3gtsF41598U1@individual.net>
Robert Maas, see http://tinyurl.com/uh3t wrote:
>> From: jayessay <······@foo.com>
>> http://seeker.dice.com/jobsearch/servlet/JobSearch?op=101&···············································@activejobs0&c=1&source=1
>
> My experience with DICE is that none of the jobs they advertise are
> legitimate in the sense of anyone actually responding to resumes sent.
> But I went ahead and tried to apply anyway:
> - They actively discriminate against anyone without the money to
> purchase MicroSoft Word.
How do you know already? Did they tell you that they don't like your
PDF or ASCII file? (well, in the second case I could understand them)
> - The only space for sending my resume is a single-line textfield, so
> the most I could put there was one URL pointing to my resume.
Maybe the textfield wasn't for arbitrary text?
> - When I clicked the submit button, I got an error message something
> like "can't open file for upload". I've never seen that particular
> error message from any other WebSite, so the Dice site seems to be
> broken in a unique way.
Hm, *maybe* the textfield is not for an URL, but for a file, and the
submit button posts the file to their server?
Hell, and I haven't even tried their application website...
> Now if you can find a recent ad for a Lisp job which doesn't
> discriminate against people who don't already have jobs with lots of
> money to purcharse MicroSoft products, and where there is actually a
> working submit-resume process, I'd be glad to try again.
Well, there's always OpenOffice (or NeoOffice) for free, but it
discriminates against people with less than 256MB memory in their
machine, or with non-mainstream operating systems.
> But thanks for trying anyway.
Good to see that there are Lisp jobs...
--
My ideal for the future is to develop a filesystem remote interface
(a la Plan 9) and then have it implemented across the Internet as
the standard rather than HTML. That would be ultimate cool.
Ken Thompson
From: [Invalid-From-Line]
Subject: Re: Any LISP programmers able to earn a living in the "real world"
Date:
Message-ID: <87br37b2r3.fsf@kafka.homenet>
·······@Yahoo.Com (Robert Maas, see http://tinyurl.com/uh3t) writes:
> - They actively discriminate against anyone without the money to
> purchase MicroSoft Word.
Either use Open Office or do what I do:
ie: Use latex and then generate rtf ( only because I like using emacs +
latex for most things )
> - When I clicked the submit button, I got an error message something
> like "can't open file for upload".
That obviously uploads a file from your machine to their server.
>so the Dice site seems to be
> broken in a unique way.
Perhaps Maas is broken is a unique way...
--
Seek simplicity and mistrust it.
Alfred Whitehead
A witty saying proves nothing.
Voltaire
·······@Yahoo.Com (Robert Maas, see http://tinyurl.com/uh3t) writes:
> > From: jayessay <······@foo.com>
> > http://seeker.dice.com/jobsearch/servlet/JobSearch?op=101&···············································@activejobs0&c=1&source=1
>
> My experience with DICE is that none of the jobs they advertise are
> legitimate in the sense of anyone actually responding to resumes sent.
> But I went ahead and tried to apply anyway:
> - They actively discriminate against anyone without the money to
> purchase MicroSoft Word.
> - The only space for sending my resume is a single-line textfield, so
> the most I could put there was one URL pointing to my resume.
> - When I clicked the submit button, I got an error message something
> like "can't open file for upload". I've never seen that particular
> error message from any other WebSite, so the Dice site seems to be
> broken in a unique way.
None of this is relevant to your original claim, which clearly is
false. Presumably if you add enough qualifications you can end up
saying that there have not even been any Java jobs offered during the
last ten years. Shrug.
/Jon
--
'j' - a n t h o n y at romeo/charley/november com