Hi
I'm trying to understand what it is that people like about LISP and whether
I should learn it. (I'm interested in AI)
Could somebody offer an example of a short (a few Kb or less) program in
LISP that does something useful and that would be relatively hard and
time-consuming to program in C++? (aside from such built-in niceties as
arbitrary-precision arithmetic please)
Thanks
Wroot
--
Please note: mail sent to ·····@my-deja.com is ignored.
Wroot <·····@my-deja.com> wrote in message news:<············@newsmaster.cc.columbia.edu>...
> Could somebody offer an example of a short (a few Kb or less) program in
> LISP that does something useful and that would be relatively hard and
> time-consuming to program in C++? (aside from such built-in niceties as
Give us an example of the kind of program you mean, by posting the C++
version of it. Be sure it has clear enough comments that a non-C++
programmer can understand what it's for, what it does, how it works,
etc. It will be interesting to see how many Lisp versions of it you
get in response, and how much shorter and neater most of them are than
the C++ version.
Wroot wrote:
> Could somebody offer an example of a short (a few Kb or less) program in
> LISP that does something useful and that would be relatively hard and
> time-consuming to program in C++? (aside from such built-in niceties as
> arbitrary-precision arithmetic please)
How about the Othello program from Norvig's PAIP. having programmed
alpha-beta search, transposition tables, and `board abstractions in both
Lisp and C++, I'd say Lisp is much nicer...
faa
Frank A. Adrian wrote:
> Wroot wrote:
>
>> Could somebody offer an example of a short (a few Kb or less) program in
>> LISP that does something useful and that would be relatively hard and
>> time-consuming to program in C++? (aside from such built-in niceties as
>> arbitrary-precision arithmetic please)
>
> How about the Othello program from Norvig's PAIP. having programmed
> alpha-beta search, transposition tables, and `board abstractions in both
> Lisp and C++, I'd say Lisp is much nicer...
>
> faa
I downloaded http://www.norvig.com/paip/othello.lisp
It compiled (clisp -c othello.lisp), but couldn't load:
clisp_prompt>> (load "othello")
Wroot
--
Please note: mail sent to ·····@my-deja.com is ignored.
Wroot <·····@my-deja.com> writes:
> Frank A. Adrian wrote:
>
> > Wroot wrote:
> >
> >> Could somebody offer an example of a short (a few Kb or less) program in
> >> LISP that does something useful and that would be relatively hard and
> >> time-consuming to program in C++? (aside from such built-in niceties as
> >> arbitrary-precision arithmetic please)
> >
> > How about the Othello program from Norvig's PAIP. having programmed
> > alpha-beta search, transposition tables, and `board abstractions in both
> > Lisp and C++, I'd say Lisp is much nicer...
> >
> > faa
>
> I downloaded http://www.norvig.com/paip/othello.lisp
> It compiled (clisp -c othello.lisp), but couldn't load:
> clisp_prompt>> (load "othello")
A little lesson on debugging Lisp:
[1]> (load "othello.lisp")
;; Loading file othello.lisp ...
*** - FUNCTION: undefined function SYMBOL
1. Break [2]> backtrace
EVAL frame for form #'SYMBOL
EVAL frame for form (CROSS-PRODUCT #'SYMBOL '(? A B C D E F G H ?) '(? 1 2 3 4 5 6 7 8 ?))
EVAL frame for form
(LET
((SQUARE-NAMES
(CROSS-PRODUCT #'SYMBOL '(? A B C D E F G H ?) '(? 1 2 3 4 5 6 7 8 ?))
))
(DEFUN H8->88 (STR) "Convert from alphanumeric to numeric square notation."
(OR (POSITION (STRING STR) SQUARE-NAMES :TEST #'STRING-EQUAL) STR)
)
(DEFUN 88->H8 (NUM) "Convert from numeric to alphanumeric square notation."
(IF (VALID-P NUM) (ELT SQUARE-NAMES NUM) NUM)
) )
EVAL frame for form (LOAD "othello.lisp")
Huh, he seems to think there's a function named SYMBOL. If I'd read
the book, maybe I'd understand why (*ahem*). But let's try this:
[···@malatesta foo]$ grep -rni "defun symbol" *.lisp
auxfns.lisp:222:(defun symbol (&rest args)
Okay, so it looks like he's got some auxillary functions. Let's load
them then try to finish loading othello.lisp:
1. Break [2]> (load "auxfns.lisp")
;; Loading file auxfns.lisp ...
;; Loading of file auxfns.lisp is finished.
T
1. Break [2]> redo
;; Loading of file othello.lisp is finished.
T
However, I imagine you'll get a lot more from it if you read the
chapter in the book along with playing with the code. If you live
near a good library, they'll likeley have it (note the slightly
tautological usage of the word "good" ;)
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
From: Erik Naggum
Subject: Re: realistic but short and simple LISP examples?
Date:
Message-ID: <3217434489399498@naggum.net>
* Wroot <·····@my-deja.com>
| I'm trying to understand what it is that people like about LISP and
| whether I should learn it. (I'm interested in AI)
After you have written and debugged and deployed several hundred thousand
lines of code in languages X, Y, and Z, and then find yourself shouting
"there's _got_ to be a better way than this!", Common Lisp replies "you
called, master?" What I like about Common Lisp is that it takes care of
90% of the boring details for me that I have had to deal with in every
other language I have used. Lisp is all about this delivery from tedium.
(Hence, Scheme is not a Lisp in my view.)
Incidentally, "LISP" is generally taken to be the pre-1980 typewritten
manuscript rendition of the small caps tradition that was used to print
acronyms nicely and without making them stand out like sore thumbs from
the page. This tradition has largely been superseded by initial caps and
the rest lowercased. So what an HTML-based author would write as
<small>LISP</small> is now just Lisp. You date yourself (or your
sources) back at least 20 years when you write it in all caps. Even
Fortran and Cobol are written like words, now.
| Could somebody offer an example of a short (a few Kb or less) program in
| LISP that does something useful and that would be relatively hard and
| time-consuming to program in C++? (aside from such built-in niceties as
| arbitrary-precision arithmetic please)
The real niceties are invisible to C++ programmers, because they would no
more look for them than an illeterate would consult encyclopedia to learn.
It is impossible to see certain things until you have been alerted to
their existence, and such alerts cause too much cognitive dissonance
among people who do not believe it can exist to work without experience.
The way to get into this is just to read textbooks and the specification
and to experiment with unexpected concepts. Small examples will largely
not work, because most of the features work best with larger projects.
Anybody can create small nice things. Creating large nice things takes a
different approach entirely. Common Lisp excels at the latter, sometimes
regarded as a problem since most programmers start off with exceptionally
small things and think they can grow a small cottage into a skyscraper
simply by scaling it up and adding stuff to it.
This is not to say that Common Lisp does not have its share of small
programs that do useful things, but every time someone tries to answer
such a question, C++ people crawl out of the woodwork to show that they
can do it, too, like some paranoid and hypercompetitive athlete afraid
that someone, somewhere, is good at something. Language comparisons are
largely only annoying. Better learn the new language. Yes, you should
learn Common Lisp. (Do not bother wasting your time on Scheme.)
///
--
The past is not more important than the future, despite what your culture
has taught you. Your future observations, conclusions, and beliefs are
more important to you than those in your past ever will be. The world is
changing so fast the balance between the past and the future has shifted.
Erik Naggum wrote:
> Small examples will largely
> not work, because most of the features work best with larger projects.
> Anybody can create small nice things. Creating large nice things takes
> a
> different approach entirely. Common Lisp excels at the latter,
> sometimes regarded as a problem since most programmers start off with
> exceptionally small things and think they can grow a small cottage into
> a skyscraper simply by scaling it up and adding stuff to it.
C++ was really created for large projects and to facilitate code reuse too,
and as its creator said, "For code to be reusable, it first has to be
usable". So I don't really buy this argument. Sorry. Besides, small is
beautiful.
Wroot
--
Please note: mail sent to ·····@my-deja.com is ignored.
From: Erik Naggum
Subject: Re: realistic but short and simple LISP examples?
Date:
Message-ID: <3217465609454072@naggum.net>
* Wroot <·····@my-deja.com>
| C++ was really created for large projects and to facilitate code reuse
| too, and as its creator said, "For code to be reusable, it first has to
| be usable". So I don't really buy this argument. Sorry.
Well, I basically said as much. You are facing some counter-information
to your previously held beliefs and your reaction is known as cognitive
dissonance. Unless trained to deal with this feeling, it appears from
the literature that "most people" will reject the counter-information out
of hand, rather than investigate _why_ they think it is wrong. I tend to
regard contradictions and counter-information as a signal to re-evaluate
what I thought I knew to be correct, and try to understand how both may
be correct at the same time because they exist in different contexts,
thus perhaps telling me something important I did not know previously.
Since there are also a large number of idiots out there who have no idea
how to deal with either information or counter-information, but instead
believe only that which causes _no_ cognitive dissonance, which applies
to the _first_ piece of information in any suitably narrowly defined
area, or that which some suitably irrational authority tells them, there
will be a lot of people who produce counter-factual disinformation which
looks very much like counter-information, thus wasting people's time with
their craploads of nonsense. This is extremely annoying, and the habit
of "most people" to ignore that which causes cognitive dissonance may be
a good defense mechanism against the usual crapload of desinformation
that we are exposed to, but it is not a good thing for an engineer or a
scientist or an investigator of any sort of truth.
Take a good look at my .signature and try to accept that what you have
yet to learn will always out-volume what you know. It is therefore much
more important to determine whether what you already know is false than
to "confirm" that it is true or just to assume that it is.
C++ fails at its goals. This is hard for C++ programmers to accept, but
evident to everybody else. Common Lisp succeeds at its goals. This is
harder for C++ programmers to accept, but unfortunately evident only to
past and present Common Lisp programmers. As life is too long to become
an expert at C++, so those who have invested the effort tend to want what
they believe to be true, regardless of the effort required to make it so.
Learn Common Lisp. Get into its spirit. It will change you, so you have
to let it. Resist the temptation to write C++ in Common Lisp. Absorb
the language and the environments you get access to. Start with one of
the commercial offerings. I recommend Allegro CL from Franz Inc
(www.franz.com), but others think others would recommend LispWorks from
Xanalys (www.xanalys.com). The commercial offerings are so much better
for learning the language than the free offerings, which currently lack
good development environments.
| Besides, small is beautiful.
It is actually the other way around: Anything is beautiful when small.
The hard part is making something larger be beautiful. This is where
most languages _really_ suck, which is why they want small examples.
Please think about the analogy to building a skyscraper by starting from
a small cottage. Or think about scaling an ant up to the size of an
cougar (one of the most beautiful animals I know), or an elephant.
///
--
The past is not more important than the future, despite what your culture
has taught you. Your future observations, conclusions, and beliefs are
more important to you than those in your past ever will be. The world is
changing so fast the balance between the past and the future has shifted.
> * Wroot <·····@my-deja.com> tells us:
> | C++ was really created for large projects and to facilitate code reuse
> | too,
To which Erik Naggum <····@naggum.net> responds:
> C++ fails at its goals.
This, of course, is the unvarnished truth. Indeed, the emperor _has_
_no_ _clothes_.
Wroot, I wonder, have you _worked_ on a large project in C++? Or even
just a medium sized one? The largest project I worked on in C++ was
in the 500,000 LOC range (20 programmers for 18 months, and would
be considered "medium size" in the industry.)
It was _hell_. Hour long compiles, straight-jacket design to insure
interfaces don't break every day, mind-numbing debug sessions finding
out where some moron^H^H^H^H^H lazy contractor was obliviously trashing
the stack... well, I could go on, but you get the point.
I now believe that that project could have been completed, more successfully
(give C++ credit, we succeeded!) with 4-6 good lispers in < 12 months.
HOW in the world is C++ helping large projects and re-use? I'll give
you a hint: there's no tool vendor selling Purify(TM) for CL. In fact,
there's an entire _slew_ of those bogus "helper" tools aimed at the C++
market (e.g. http://www.parasoft.com/). You'd think _somebody_ would have
figured out that if you need all this extra help, perhaps there's a better
way? But, oh, I remember, that involves _thinking_. Pass.
That crowd is _finally_ moving to Java, so there's hope! In a decade,
they may yet learn about macros. In two decades, they may discover
the MOP. Meantime, take my word for it: life is just _too_ _short_.
Learn CL, you won't regret it.
--
It would be difficult to construe Larry Wall, in article
this as a feature. <·····················@netlabs.com>
Alain Picard wrote:
> The largest project I worked on in C++ was
> in the 500,000 LOC range (20 programmers for 18 months, and would
> be considered "medium size" in the industry.)
That comes out to 8 lines per man-hour. Weren't you guys about 8 times
overworked by industry standards?
> I now believe that that project could have been completed, more
> successfully (give C++ credit, we succeeded!) with 4-6 good lispers in <
> 12 months.
What was the nature of the project?
> there's no tool vendor selling Purify(TM) for CL.
The lack of "helper" software for Lisp does not prove its superiority to
languages that do create markets for such software.
> That crowd is _finally_ moving to Java, so there's hope! In a decade,
> they may yet learn about macros. In two decades, they may discover
> the MOP. Meantime, take my word for it: life is just _too_ _short_.
>
> Learn CL, you won't regret it.
Thanks for your advice
Wroot
--
Please note: mail sent to ·····@my-deja.com is ignored.
Wroot <·····@my-deja.com> writes:
> What was the nature of the project?
It was a service switching point, which is an ITU standardized
(at the protocol level) component of an Intelligent Network.
So, basically a telco app; an add-on to a regular switch to
analyze calls and re-route some of them based on certain criteria.
> > there's no tool vendor selling Purify(TM) for CL.
>
> The lack of "helper" software for Lisp does not prove its superiority to
> languages that do create markets for such software.
Of course it does. Please think about what you are saying.
> >
> > Learn CL, you won't regret it.
>
> Thanks for your advice
Well, it was well intentioned. I hope you receive it that way.
Would you do me the courtesy of telling me what is the largest
C++ project you have personally worked on? Thanks.
--
It would be difficult to construe Larry Wall, in article
this as a feature. <·····················@netlabs.com>
Alain Picard wrote:
> Wroot <·····@my-deja.com> writes:
>> The lack of "helper" software for Lisp does not prove its superiority to
>> languages that do create markets for such software.
>
> Of course it does. Please think about what you are saying.
There is a language called D which was supposed to succeed C. Do you know
of any "helper" software for it? No? By your logic, we should now say that
D is superiour to C++ and many others.
> Would you do me the courtesy of telling me what is the largest
> C++ project you have personally worked on? Thanks.
7,500
Wroot
--
Please note: mail sent to ·····@my-deja.com is ignored.
Wroot <·····@my-deja.com> wrote in message news:<············@newsmaster.cc.columbia.edu>...
> Alain Picard wrote:
>
> > Wroot <·····@my-deja.com> writes:
> >> The lack of "helper" software for Lisp does not prove its superiority to
> >> languages that do create markets for such software.
> >
> > Of course it does. Please think about what you are saying.
>
> There is a language called D which was supposed to succeed C. Do you know
> of any "helper" software for it? No? By your logic, we should now say that
> D is superiour to C++ and many others.
You're starting to sound like a troll.
There are thousands of Common Lisp programmers, using at least 4 high
quality commercial Common Lisp IDEs, and a number of free compilers.
There *is* a market for tools here. The fact that there are no
"helper" products follows from the fact that they are not needed,
because the language (Common Lisp) is not broken at the design level
in the way that C++ is.
In contrast, if the painfully obvious still needs to be pointed out,
there are no helper tools for D, because practically no one uses D,
there are not 3 thriving Windows implementations of D (as there are
for Common Lisp), there is no high quality commercial D IDE for the
Mac or Linux (as there are for Common Lisp), there is no comp.lang.d
newsgroup, etc. etc.
D is dead, but Common Lisp is quite alive, and commercially viable.
Therefore, the absence of "helper" tools for Common Lisp means they
are unnecessary because the language works the way it's design
community intended, unlike C++.
From: Andreas Eder
Subject: Re: realistic but short and simple LISP examples?
Date:
Message-ID: <m31yhulu84.fsf@elgin.eder.de>
Wroot <·····@my-deja.com> writes:
> Alain Picard wrote:
>
> > Wroot <·····@my-deja.com> writes:
> >> The lack of "helper" software for Lisp does not prove its superiority to
> >> languages that do create markets for such software.
> >
> > Of course it does. Please think about what you are saying.
>
> There is a language called D which was supposed to succeed C. Do you know
> of any "helper" software for it? No? By your logic, we should now say that
> D is superiour to C++ and many others.
No, you've got it the wrong way. It should be:
"If it there is helper software, then it is not good" or logically
equivalent:
"If it is good. then it doesn't have helper software".
You said: "If there is no helper software, then it is good". A common
fallacy, but nevertheless wrong.
A. Eder
--
Wherever I lay my .emacs, there�s my $HOME.
Wroot <·····@my-deja.com> writes:
> Alain Picard wrote:
>
> > Wroot <·····@my-deja.com> writes:
> >> The lack of "helper" software for Lisp does not prove its superiority to
> >> languages that do create markets for such software.
> >
> > Of course it does. Please think about what you are saying.
>
> There is a language called D which was supposed to succeed C. Do you know
> of any "helper" software for it? No? By your logic, we should now say that
> D is superiour to C++ and many others.
Message id <············@newsmaster.cc.columbia.edu>:
> P.S. If I were trolling, I would have cross-posted to
> comp.lang.c++. Why would you accuse me of something like this?
Gee, I have no idea.
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
Wroot <·····@my-deja.com> writes:
> Alain Picard wrote:
>
> There is a language called D which was supposed to succeed C. Do you know
> of any "helper" software for it? No? By your logic, we should now say that
> D is superiour to C++ and many others.
Is this language in common use? Would it make commercial sense for a
company to come out with a product to help D programmers? I was trying
to give you the benefit of the doubt, but it's becoming clear that you
are just another troll.
> > Would you do me the courtesy of telling me what is the largest
> > C++ project you have personally worked on? Thanks.
>
> 7,500
Thank you. Now, I don't know where you read that C++ was designed for
large projects, but since you have not had the misfortune to wait for
multi-hour long compiles and debug subtly broken builds, let me assure you that,
if that really was a design parameter, the designer failed utterly.
BTW, there is an excellent book, by Lakos, on designing large C++ systems.
The amount of planning and care involved to succeed in such an endeavour,
and the fact that it needs to be so painstakingly explained [this is a
_very_ didactic book, thank God] _shows_ that C++ could not have been
"designed with large systems in mind".
On another note, I was shocked to read in another thread that Bjarne
didn't understand the difference between a language specification and
implementation, and that's why we now have C++. I'll now have to read
"The design and evolution of C++", just out of morbid curiosity.
Can you imagine if all these hordes of programmers had instead been
working on an efficient, souped up simula instead? Gee... MicroSoft
might have never happened. :-)
--
It would be difficult to construe Larry Wall, in article
this as a feature. <·····················@netlabs.com>
On Sat, 15 Dec 2001 18:17:13 -0500, Wroot <·····@my-deja.com> wrote:
>C++ was really created for large projects and to facilitate code reuse too,
>and as its creator said, "For code to be reusable, it first has to be
Common Lisp is much better for that purpose, specially if you are working in a
small team, trust me.
It's object system is far more powerful and MUCH easier to use. All the
generative and template meta-programming that is considered so cool and modern
in the C++ community is already old stuff in CL, and several orders of
magnitude easier to use. In lisp it's not just cool, it's _feasible_.
Take the time to learn lisp, it's a great tool to add to your toolchest. Make
sure you learn CLOS (the object system) as well as macros (templates on
steroids).
>usable". So I don't really buy this argument. Sorry. Besides, small is
>beautiful.
It trully is: Borland C++ Builder takes more than 300MB in my HD, while
Lispworks (a roughly equivalent Lisp compiler/ide) from Xanalys takes less
than 20MB. Besides, it's also slightly less expensive. Small (both in my HD as
my wallet) IS BEAUTIFUL! ;-)
--
Fernando Rodr�guez
frr at wanadoo dot es
--
From: Kenny Tilton
Subject: Re: realistic but short and simple LISP examples?
Date:
Message-ID: <3C1BCD4A.F3BE578D@nyc.rr.com>
Want me to send you the distro for Cells (ne Semaphors)?
(1) You can play with Lisp instead of reading about it (one good thing
about Lisp) and
(2) you can explore a not so new dataflow paradigm I stumbled on
independently while just doing neat stuff with Lisp (one bad thing about
using Lisp commercially is that any sufficiently complex project will
lead a Lisper to hack up a pet package and lose all interest in the
original project).
btw, I like: standard, compiled, fast, mature, interactive, closures,
macros, GC, CLOS, restarts in the backtrace, MetaObject programming,
single typespace (no Integer vs. int ala Java),special variables, all
the built-in functions, automatic formatting of code, the #$%^#%(
package system and logical naming scheme, AND THE GODDAMN PARENTHESES!
OK?!!
sorry.
:)
kenny
clinisys
Wroot wrote:
>
> Hi
>
> I'm trying to understand what it is that people like about LISP and whether
> I should learn it. (I'm interested in AI)
>
> Could somebody offer an example of a short (a few Kb or less) program in
> LISP that does something useful and that would be relatively hard and
> time-consuming to program in C++? (aside from such built-in niceties as
> arbitrary-precision arithmetic please)
>
> Thanks
>
> Wroot
> --
> Please note: mail sent to ·····@my-deja.com is ignored.
On Sat, 15 Dec 2001 02:39:29 -0500, Wroot <·····@my-deja.com> wrote:
>Hi
>
>I'm trying to understand what it is that people like about LISP and whether
>I should learn it. (I'm interested in AI)
>
>Could somebody offer an example of a short (a few Kb or less) program in
>LISP that does something useful and that would be relatively hard and
>time-consuming to program in C++? (aside from such built-in niceties as
>arbitrary-precision arithmetic please)
Get a copy of "Paradigms of artificial Intelligence" by Peter Norvig. It's
full of small and not so small great examples. It's one of the very best
programming books I've read.
--
Fernando Rodr�guez
frr at wanadoo dot es
--