Dylan (DYnamic LANguage) -- what's the deal? 
Author Message
 Dylan (DYnamic LANguage) -- what's the deal?

Dylan, huh?  Just what is going on there?
Here is why I ask.

On Sun.30.Jul.2000 I perused a paper by Peter Norvig and
David Cohn called "Adaptive Software" at the URL
http://www.*-*-*.com/

The authors seemed to make a wild and irresponsible claim
that with Dylan, it was possible for a running program to
grow and add new functionality on the fly, so to speak,
that is, without stopping the program and re-coding it.

Well, suppose you hade this artificial intelligence (AI)
program in Dylan, fashioned after the PD AI Mind.Forth at
http://www.*-*-*.com/ (q.v.).

Suppose further that you wanted your AI program to run
indefinitely for years and years, growing but not halting.
Can Dylan do the trick?  Can Dylan yield an immortal AI?

Arthur T. Murray

--
http://www.*-*-*.com/ ~mentifex : Home Page
http://www.*-*-*.com/ : AI in Dylan
http://www.*-*-*.com/ : DIY AI

Sent via Deja.com http://www.*-*-*.com/
Before you buy.



Wed, 22 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?


Quote:
> On Sun.30.Jul.2000 I perused a paper by Peter Norvig and
> David Cohn called "Adaptive Software" at the URL
> http://www.*-*-*.com/

> The authors seemed to make a wild and irresponsible claim
> that with Dylan, it was possible for a running program to
> grow and add new functionality on the fly, so to speak,
> that is, without stopping the program and re-coding it.

> Arthur T. Murray


Arthur -

The capability you refer to is entirely standard for the Lisp family of
languges, of which Dylan is a member. Far from being {*filter*}, it's a
capability
that's decades old and is shared by several other languages.

What puzzles me is that you've posted constantly (often on inappropriate
groups)
concerning AI for years, and yet you don't know such a basic fact about the
most common of all AI languages, Lisp. Even AI programmers who work in
Prolog
or C++ normally find it necessary to be able to read Lisp, so that they can
evaluate and
learn from their fellows' work.

Peter Norvig is one of the most valuable and talented members of both the AI
and computer language communities. You might want to keep
words like "wild and irresponsible" further away from his name - and indeed
anyone else's name, until you have real reason to use such terms. If you
want to learn more about AI, Lisp, or computer languages, then reading one
or more of Peter Norvig's books would be an excellent place to start.

Jonathan Coupe



Wed, 22 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?
Jonathan Coupe in gentlemanly fashion wrote on Sat, 5 Aug 2000:

Quote:


>> On Sun.30.Jul.2000 I perused a paper by Peter Norvig and
>> David Cohn called "Adaptive Software" at the URL
>> http://www.*-*-*.com/

>> The authors seemed to make a wild and irresponsible claim
>> that with Dylan, it was possible for a running program to
>> grow and add new functionality on the fly, so to speak,
>> that is, without stopping the program and re-coding it.

>> Arthur T. Murray

> Arthur -

> The capability you refer to is entirely standard for the
> Lisp family of languges, of which Dylan is a member.
> Far from being {*filter*}, it's a capability
> that's decades old and is shared by several other languages.

> What puzzles me is that you've posted constantly (often on
> inappropriate groups) concerning AI for years, and yet you
> don't know such a basic fact about the most common of all
> AI languages, Lisp.

ATM:
With my B.A. degree in ancient Greek and Latin, I can't claim
to be much of a programmer; I just barely get along with
http://www.*-*-*.com/

I guess I don't learn very many languages; I just pick up
"nuggets" about them, as I did from Norvig's and Cohen's paper.
Theory of how language may work in the brain-mind is my forte:
http://www.*-*-*.com/

Quote:
> Even AI programmers who work in Prolog or C++ normally find it
> necessary to be able to read Lisp, so that they can evaluate
> and learn from their fellows' work.

> Peter Norvig is one of the most valuable and talented members
> of both the AI and computer language communities.
> You might want to keep words like "wild and irresponsible"
> further away from his name - and indeed anyone else's name,

ATM:
Here I do have a valid excuse.  I pre-pended "seemed to"
and then I trying-to-be-witty used those outrageous words
just to express my neophyte-programmer astonishment at the
idea that a running program could go through changes.  I
hope that it was obvious to everybody that I was not maligning
Peter Norvig, just trying to show my surprise (and this is
not an after-the-facts cowardly excuse, but the truth).

Quote:
> until you have real reason to use such terms. If you want
> to learn more about AI, Lisp, or computer languages, then reading one
> or more of Peter Norvig's books would be an excellent place to start.

> Jonathan Coupe

Thank you most sincerely for your kind help.

Arthur T. Murray



Wed, 22 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?
/ What puzzles me is that you've posted constantly (often on inappropriate
/ groups)
/ concerning AI for years, and yet you don't know such a basic fact about the
/ most common of all AI languages, Lisp. Even AI programmers who work in

Because he's doing his work in Forth. Forth doesn't have close ties to
Lisp or functional programming in general.

=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
Sign up for WASHINGTON MUTUAL BANK's special
We Rob You While You Sleep Service TODAY!
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
CACS: Collective Against Consensual Sanity       v0.123
Now a text site map http://www.tsoft.com/~wyrmwif/cacs/
pretty?     http://www.geocities.com/SoHo/Studios/5079/
:)-free zone.                 Elect LUM World Dictator!



Wed, 22 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?
Centuries ago, Nostradamus foresaw a time when China Black & Blue would say:

Quote:
>/ What puzzles me is that you've posted constantly (often on inappropriate
>/ groups)
>/ concerning AI for years, and yet you don't know such a basic fact about the
>/ most common of all AI languages, Lisp. Even AI programmers who work in

>Because he's doing his work in Forth. Forth doesn't have close ties to
>Lisp or functional programming in general.

That doesn't mean that it is safe to remain ignorant of the other
languages' approaches to AI.

Doing _better_ should involve a synthesis of the understandings
coming from "Mentifex's own work" in addition to the things he might
learn from looking at other languages.

When he suggested that it was somewhat unbelievable that Dylan could
possibly support "dynamic" programming, my first thought was
"ignorant {*filter*};" Dylan is merely the most recent of the Lisp Language
Family, and the old and rich history is one where a major point is the
support of introspective programming.

And actually, these days, languages with the EVAL function are getting
to be a dime a dozen, as the scripting languages like Perl and python
that have grown quite popular also support introspection.
--

"Free software: the Source will be with you, always."



Thu, 23 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?

Quote:

>And actually, these days, languages with the EVAL function are getting
>to be a dime a dozen, as the scripting languages like Perl and Python
>that have grown quite popular also support introspection.

Maybe by 2010, "mainstream" programming languages will catch
up to c. 1965 Lisp.  With any luck, by 2020, they'll manage to get
to c. 1985 Lisp and Smalltalk.


Thu, 23 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?
As I've said before: in the same way that all operating systems want to be
UNIX, all programming languages want to be Lisp. This makes Dylan the MacOS
X or Gnome of the Lisp family. :-)

- Rob.

Quote:

> Organization: Road Runner
> Newsgroups:
> comp.lang.dylan,comp.lang.misc,comp.ai.philosophy,comp.software-eng
> Date: Sun, 06 Aug 2000 13:20:43 GMT
> Subject: Re: Dylan (DYnamic LANguage) -- what's the deal?


>> And actually, these days, languages with the EVAL function are getting
>> to be a dime a dozen, as the scripting languages like Perl and Python
>> that have grown quite popular also support introspection.

> Maybe by 2010, "mainstream" programming languages will catch
> up to c. 1965 Lisp.  With any luck, by 2020, they'll manage to get
> to c. 1985 Lisp and SmallTalk.



Thu, 23 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?
/ Maybe by 2010, "mainstream" programming languages will catch
/ up to c. 1965 Lisp.  With any luck, by 2020, they'll manage to get
/ to c. 1985 Lisp and SmallTalk.

And maybe by 2020, Lisp and Smalltalk will be efficient enough to write
operating system kernels and interrupt drivers in.

In the Year 2525....

=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
Sign up for WASHINGTON MUTUAL BANK's special
We Rob You While You Sleep Service TODAY!
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
CACS: Collective Against Consensual Sanity       v0.123
Now a text site map http://www.tsoft.com/~wyrmwif/cacs/
pretty?     http://www.geocities.com/SoHo/Studios/5079/
:)-free zone.                 Elect LUM World Dictator!



Thu, 23 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?


Quote:
>/ Maybe by 2010, "mainstream" programming languages will catch
>/ up to c. 1965 Lisp.  With any luck, by 2020, they'll manage to get
>/ to c. 1985 Lisp and SmallTalk.

>And maybe by 2020, Lisp and Smalltalk will be efficient enough to write
>operating system kernels and interrupt drivers in.

By 1980, the Lisp Machine OS was written entirely in Lisp.
And, by 1985, the Lisp Machine's front-end processor --
a 68000 -- had a proprietary OS which was also entirely
written in a dialect of Lisp called LIL, which was just Lisp
with type declarations.  Both of these OSes had their
device drivers, interrupt code, virtual paging system, etc,
entirely written in Lisp.  I can speak on this authoritatively,
since I was one of the implementors.  The performance
of my c. 1990 MacIvory III -- a pathetically slow machine --
is still very usable, even compared to modern machines.

FWIW, Dylan benchmarks for Functional Dylan (nee Harlequin
Dylan) were regularly competetive and sometimes beat
their counterparts written in C.  And that was a compiler
written by a comparatively tiny team.  And Harlequin's and
Franz's Lisp implementations often did better tan Dylan.

As I said, maybe by 2020, the "mainstream" world will
catch up with the state of the art that existed 35 years earlier.



Thu, 23 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?
/ By 1980, the Lisp Machine OS was written entirely in Lisp.
/ And, by 1985, the Lisp Machine's front-end processor --

A real commercial success.

=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
Sign up for WASHINGTON MUTUAL BANK's special
We Rob You While You Sleep Service TODAY!
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
CACS: Collective Against Consensual Sanity       v0.123
Now a text site map http://www.tsoft.com/~wyrmwif/cacs/
pretty?     http://www.geocities.com/SoHo/Studios/5079/
:)-free zone.                 Elect LUM World Dictator!



Thu, 23 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?
You can't judge something's cultural/historical value or its eventual
impact by its immediate commercial success. Take Microsoft:  please.
With services and device drivers writable in J*v*, and the next version
of C++ looking to add GC to go with the generic programming it's
recently had tacked on, it's plain that the bleating "mainstream" is
playing a slow game of catch-up with forty years of Lisp achievements.
And, no, I'm not a Lisp booster by trade.

- Rob.


Quote:

> / By 1980, the Lisp Machine OS was written entirely in Lisp.
> / And, by 1985, the Lisp Machine's front-end processor --

> A real commercial success.

--
Rob Myers - http://www.robmyers.org/   H2G2 - http://www.h2g2.com/
MacOS wonderfulness for The Hitch Hiker's Guide to the Galaxy Game.
"Don't talk to sociologists. Social practice has no sociological
content." - Art & Language.


Fri, 24 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?

Quote:

> You can't judge something's cultural/historical value or its eventual
> impact by its immediate commercial success. Take Microsoft:  please.
> With services and device drivers writable in J*v*, and the next version
> of C++ looking to add GC to go with the generic programming it's
> recently had tacked on, it's plain that the bleating "mainstream" is
> playing a slow game of catch-up with forty years of Lisp achievements.
> And, no, I'm not a Lisp booster by trade.

At that time these Lisp systems were a commercial success.
The Lisp companies sold hardware/software for several hundred
million dollars over the years.

Quote:

> - Rob.


> > / By 1980, the Lisp Machine OS was written entirely in Lisp.
> > / And, by 1985, the Lisp Machine's front-end processor --

> > A real commercial success.

--
Rainer Joswig, Hamburg, Germany



Fri, 24 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?


Quote:

> / By 1980, the Lisp Machine OS was written entirely in Lisp.
> / And, by 1985, the Lisp Machine's front-end processor --

> A real commercial success.

You said

Quote:
> And maybe by 2020, Lisp and Smalltalk will be efficient enough to write
> operating system kernels and interrupt drivers in.

Scott McKay clearly said it could be and had been done, and was
technically successfull.

Commercial success wasn't mentioned by you initially.

Do you claim technically good = commercially successful? If so, what
sort of technical attributes must a language/system have that ensures
commercial viability?

Do you claim technical cruddism will necessarily = failure? OS-wise I'm
thinking MS Win 3.1

jt

Quote:

> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
> Sign up for WASHINGTON MUTUAL BANK's special
> We Rob You While You Sleep Service TODAY!
> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
> CACS: Collective Against Consensual Sanity       v0.123
> Now a text site map http://www.tsoft.com/~wyrmwif/cacs/
> pretty?     http://www.geocities.com/SoHo/Studios/5079/
> :)-free zone.                 Elect LUM World Dictator!



Fri, 24 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?


Quote:
>/ By 1980, the Lisp Machine OS was written entirely in Lisp.
>/ And, by 1985, the Lisp Machine's front-end processor --

>A real commercial success.

What's your point?  That commercial success is an indicator
of excellent technology?  So explain why Apollo failed, despite
its high level of excellent technology.  Explain why Microsoft
succeeded in the face of its {*filter*}technology.  Commercial
success and good technology have nothing to do with each
other.  FWIW, during the 80's, Symbolics was a $100M/year
company, which by most measures is hardly a failure.

I should mention some other technologies, BTW, just to be
fair:
 - Dave Thomas, out of I-don't-remember-where in Canada,
   was building high-performance Smalltalk apps that got
   blown onto PROMs and delivered in embeddable systems
   by the early 90's at the very latest.
 - Both Symbolics and Harlequin built real-time Lisps systems
   for use in high-performance hardware switches, Symbolics
   on top of its own Lisp OS called Minima, Harlequin on top
   of VXWorks.  These systems had to meet stringent real-time
   constraints -- and did meet them.
 - Gensym has been building real-time Lisp systems for at least
   a decade, and still runs a successful business.

Your arguments just don't hold water.



Fri, 24 Jan 2003 03:00:00 GMT  
 Dylan (DYnamic LANguage) -- what's the deal?

Quote:

> The capability you refer to is entirely standard for the Lisp family of
> languges, of which Dylan is a member. Far from being {*filter*}, it's a
> capability that's decades old and is shared by several other languages.

Including Forth and its descendent postscript.

Quote:
> What puzzles me is that you've posted constantly (often on inappropriate
> groups) concerning AI for years, and yet you don't know such a basic fact
> about the most common of all AI languages, Lisp.

Why is that puzzling? I'm not that well versed in AI, but everything I've
seen him write on any subject I *do* know about has displayed similar levels
of understanding.

--
 `-_-'   In hoc signo hack, Peter da Silva.
  'U`    "Ar rug t barrg ar do mhactre inniu?"

         Disclaimer: Matthew 10:16.



Fri, 24 Jan 2003 03:00:00 GMT  
 
 [ 25 post ]  Go to page: [1] [2]

 Relevant Pages 

1. Dylan (DYnamic LANguage) -- what's the deal?

2. Dynamic Language in Esperanto (was: Re: Alleging name theft, Bob Dylan sues Apple Computer)

3. Dynamic re-definition: what they're saying on the Dylan list

4. (fwd) harlequin's dylan-corba mailing list switching to MIT's info-dylan

5. Dealing with Magnus Kempe (was Dealing with hecklers)

6. Example of dynamic loading of Dylan libraries available

7. substitution can't deal with multibyte?

8. PD OOP library - what's the deal?

9. Apple's Dylan's FAQ postable?

10. The 'Dylan' Logo

11. How to pronounce 'Dylan'?

12. How to deal with Don't cares

 

 
Powered by phpBB® Forum Software