From: ···@sef-pmax.slisp.cs.cmu.edu
Subject: Re: Future of CMU Common Lisp
Date:
Message-ID: <CB5uur.8GE.1@cs.cmu.edu>
From: ····@festival.ed.ac.uk (J W Dalton)
Why not do an innovative new software development environment
based on Common Lisp?
Common Lisp is an excellent vehicle for evolutionary programming, which I
define as any programming where you don't know exactly and completely what
you want before the code is running. (Most software development has this
evolutionary character, whether we plan it that way or not.) But Common
Lisp has some specific technical problems that make it hard to deliver
small, very fast runtime modules. One could do this by deviating in
various ways from the CL standard, but if we do that we may as well go all
the way to a new language that we like better, and that has concern for
efficient delivery built in from the start.
Perhaps more important than the real problems with Common Lisp are the
perceived ones. A lot of people out in "the mainstream" think that they
know all about Lisp, and they think that they hate it. Their view of Lisp
may be incorrect or out of date, but the battle is lost in any case. With
a new language that resembles Lisp in some ways but not in others, these
people might be persuaded to take a second look.
Is Dylan viewed as a hot area of research?
Not per se. But the problem of how to create an excellent environment for
evolutionary software development that can also produce practical,
industrial strength delivery code is a "hot" topic. It's a problem that
must be solved, and we intend to show that there are some advantages in
coming at this problem from the dynamic-language side. Lots of other
groups are starting with C or C++ and trying to make it more flexible.
-- Scott
===========================================================================
Scott E. Fahlman Internet: ····@cs.cmu.edu
Senior Research Scientist Phone: 412 268-2575
School of Computer Science Fax: 412 681-5739
Carnegie Mellon University Latitude: 40:26:33 N
5000 Forbes Avenue Longitude: 79:56:48 W
Pittsburgh, PA 15213
===========================================================================
···@sef-pmax.slisp.cs.cmu.edu writes:
> From: ····@festival.ed.ac.uk (J W Dalton)
>
> Why not do an innovative new software development environment
> based on Common Lisp?
>
>Common Lisp is an excellent vehicle for evolutionary programming, [...]
>But Common Lisp has some specific technical problems that make it hard
>to deliver small, very fast runtime modules. [...]
>Perhaps more important than the real problems with Common Lisp are the
>perceived ones. A lot of people out in "the mainstream" think that they
>know all about Lisp, and they think that they hate it. Their view of Lisp
>may be incorrect or out of date, but the battle is lost in any case. With
>a new language that resembles Lisp in some ways but not in others, these
>people might be persuaded to take a second look.
Humm. I see your point, but Dylan doesn't just resemble Lisp: right
now, it's a dialect _of_ Lisp. Even the addition of a different syntax
wouldn't change that (think of RLisp in PSL, the M-notation in the
Lisp 1.5 book, and Lisp 2, for instance). A dynamic language that
merely resembled Lisp might still be preferable to C++, but I don't
think I'd prefer it to Lisp. I'm worried that abandoning CMU CL
will make things worse for the sort of dynamic language I want rather
than better, unless someone else keeps CMU CL going.
-- jeff
In article <··········@festival.ed.ac.uk>, ····@festival.ed.ac.uk (J W Dalton) writes:
|> ···@sef-pmax.slisp.cs.cmu.edu writes:
|>
|> > From: ····@festival.ed.ac.uk (J W Dalton)
|> >
|> > Why not do an innovative new software development environment
|> > based on Common Lisp?
|> >
|> >Common Lisp is an excellent vehicle for evolutionary programming, [...]
|> >But Common Lisp has some specific technical problems that make it hard
|> >to deliver small, very fast runtime modules. [...]
|>
|> >Perhaps more important than the real problems with Common Lisp are the
|> >perceived ones. A lot of people out in "the mainstream" think that they
|> >know all about Lisp, and they think that they hate it. Their view of Lisp
|> >may be incorrect or out of date, but the battle is lost in any case. With
|> >a new language that resembles Lisp in some ways but not in others, these
|> >people might be persuaded to take a second look.
|>
|> Humm. I see your point, but Dylan doesn't just resemble Lisp: right
|> now, it's a dialect _of_ Lisp. Even the addition of a different syntax
|> wouldn't change that (think of RLisp in PSL, the M-notation in the
|> Lisp 1.5 book, and Lisp 2, for instance). A dynamic language that
|> merely resembled Lisp might still be preferable to C++, but I don't
|> think I'd prefer it to Lisp. I'm worried that abandoning CMU CL
|> will make things worse for the sort of dynamic language I want rather
|> than better, unless someone else keeps CMU CL going.
|>
|> -- jeff
I agree that Dylan will be seen as a Lisp, with all the (perceived)
negative associations, because the syntax uses parenthesis.
Thus, I don't believe you gain more than you lose by
using a new, incompatible Lisp dialect.
Why not use Scheme if CL is a lost cause?
I do think a different syntax would make a difference.
Something that looks like C, but is really Lisp. I have worked on a
language I call CRISP, which is essentially this idea.
Have a CRISP->CLOS translator,
which is essentially a C->CL translator that doesn't support C pointers.
(The C translator actually did support them, but CRISP doesn't have them).
The path of enlightment to Lisp is providing real macros at the lisp level.
Thus, in addition to supporting CPP as a first pass,
the last pass will invoke Lisp-level macros, which can be defined by
programmers. You must have good CRISP-level source code debugging for
it to really catch-on.
But ultimately I believe the language the operating system is written is
the one which people will use. Didn't DOS people code in ASSEMBLY?
Didn't the MAC people program in PASCAL??
"Write the OS in Lisp, and they will come"
-Kelly
Date: Tue, 3 Aug 1993 16:00 EDT
From: Kelly Murray <···@prl.ufl.edu>
"Write the OS in Lisp, and they will come"
The OS on Lisp Machines is written in Lisp, and nobody came.
In article <····················@SUMMER.SCRC.Symbolics.COM>, ···@stony-brook.scrc.symbolics.com (Scott McKay) writes:
|> Date: Tue, 3 Aug 1993 16:00 EDT
|> From: Kelly Murray <···@prl.ufl.edu>
|>
|> "Write the OS in Lisp, and they will come"
|>
|> The OS on Lisp Machines is written in Lisp, and nobody came.
So, "Write the *portable*, fast, robust OS in Lisp, and they will come"
That was Symbolics major strategic mistake--they thought they would
win as a hardware company, and passed up the chance to compete with
Unix/DOS/etc. as a software powerhouse. (Imagine if Symbolics were in
Microsoft's position today. Sigh...)
--
James McDonald
Kestrel Institute ········@kestrel.edu
3260 Hillview Ave. (415) 493-6871 ext. 339
Palo Alto, CA 94304 fax: (415) 424-1807
--
James McDonald
Kestrel Institute ········@kestrel.edu
3260 Hillview Ave. (415) 493-6871 ext. 339
Palo Alto, CA 94304 fax: (415) 424-1807
>>>>> "SM" == Scott McKay writes:
SM> The OS on Lisp Machines is written in Lisp, and nobody came.
We came, but then we left again (under intense pressure from
management and customers.)
- Jim
What I'd like to see when it comes to "Dynamic Programming" is a full
CLtL2 (or higher if available) with an added ability to produce small
run-time applications which would have loaded only the stuff that is
going to be used within the application.
I may be wrong about the following, but I have an impression that
Allegro CL can do something of the kind on microcomputers, e.g.,
Apple, and Windows.
I gather that the real issue is that it may be difficult to get money
from DARPA for yet another CL project, but if you replace the word "CL"
with the word "Dylan", while still keeping all the brackets, you will have
more chance to milk the cow.
--
Zdzislaw Meglicki, ·················@anu.edu.au,
Automated Reasoning Program - CISR, and Plasma Theory Group - RSPhysSE,
The Australian National University, G.P.O. Box 4, Canberra, A.C.T., 2601,
Australia, fax: (Australia)-6-249-0747, tel: (Australia)-6-249-0158
In article <··········@festival.ed.ac.uk>, ····@festival.ed.ac.uk (J W Dalton) writes:
|> Humm. I see your point, but Dylan doesn't just resemble Lisp: right
|> now, it's a dialect _of_ Lisp.
If you define "dialect _of_ Lisp" by "having lots of parens", you are
perhaps right.
But, as far as I can tell, Dylan deviates pretty strongly in some areas
from existing Lisps, and more changes and additions in areas such as
static type checking are possible. To me, that makes it a different
language from Lisp, despite its use of sexprs
To put it differently, the differences between Dylan and CL/CLOS,
Eulisp, or Scheme seem no smaller than the differences between Oberon
and Ada, which are called "different languages". And in the same way,
I would consider Lisp 1.5, MACLISP, and CommonLisp "different
languages".
|> I'm worried that abandoning CMU CL
|> will make things worse for the sort of dynamic language I want rather
|> than better, unless someone else keeps CMU CL going.
I agree that keeping CMU CL going would be nice (after all, I make a
lot of use of it and I like it better than any commercial CL because of
its type inference and because I can patch things that I consider
broken). But they aren't getting any funding anymore. On the other
hand, they aren't abandoning it either. We'll just have to see how
things shake out.
Thomas.
···@arolla.idiap.ch (Thomas M. Breuel) writes:
>In article <··········@festival.ed.ac.uk>, ····@festival.ed.ac.uk (J W Dalton) writes:
>|> Humm. I see your point, but Dylan doesn't just resemble Lisp: right
>|> now, it's a dialect _of_ Lisp.
>If you define "dialect _of_ Lisp" by "having lots of parens", you are
>perhaps right.
But I don't define "dialect of Lisp" that way, as my remark about
alternative syntax should have made clear. Nonetheless, I'm right
about Dylan being a dialect of Lisp.
>But, as far as I can tell, Dylan deviates pretty strongly in some areas
>from existing Lisps, and more changes and additions in areas such as
>static type checking are possible. To me, that makes it a different
>language from Lisp, despite its use of sexprs
There's not much difference between Dylan and:
* EuLisp level 0
* ISLisp (potential ISO standard)
* Scheme + a subset of CLOS
The strongest difference I can find is that Dylan lacks standard I/O,
something not seen since Algol 60 but reasonable if you want to do the
language first, libraries later.
Static type checking is certainly of interest to the EuLisp group.
Does the possibility of changes and additions in that direction mean
EuLisp is not a dialect of Lisp? Of course not. People want to
see Dylan as different, so they do. The name helps, of course.
People are used to thinking of langauges with different names as
different languages. People have also claimed Scheme is not Lisp.
Of course, the determined few will claim that, say, Common Lisp
is not Lisp too, but this is seldom taken seriously.
N.B. by "dialect", I am using the usual standard for "dialect"
that are applied to Lisp; they are different from those applied
to other language groups. By Lisp standard, Pascal is a dialect
of Algol, not a separate language, for instance. But if you want
to substite "language in the Lisp family" or "Lisp-like language"
for "dialect of Lisp", that's fine with me. But this being
comp.lang.lisp, I assumed I was dealing with people who would
not misunderstand me. [N.B. I wrote this before reading your
next paragraph, not expecting _you_ to take this position.]
>To put it differently, the differences between Dylan and CL/CLOS,
>Eulisp, or Scheme seem no smaller than the differences between Oberon
>and Ada, which are called "different languages". And in the same way,
>I would consider Lisp 1.5, MACLISP, and CommonLisp "different
>languages".
So? Lisp 1.5, MacLisp, and Common Lisp are all dialects of Lisp.
This is not normally considered controversial, although politics
sometimes makes it so. The same issues came up when we were
considering ISO standards, because politics made the language /
dialect difference something people could exploit. (I didn't
accept it then either.)
Anyway, whenever anyone says something like "Dylan resembles Lisp"
they're using Lisp in the generic sense. If you take the all Lisps
are different languages line, there's no language Lisp at all and
we can just as well say "Common Lisp resembles Lisp".
In short, I will not be impressed by technical quibbles about
the difference between a language and a dialect, especially when
people are perfectly willing to accept generic uses of "Lisp".
-- jd
From: Barry Margolin
Subject: Dialects vs languages (was Re: Future of CMU Common Lisp)
Date:
Message-ID: <23p3e9INN6kt@early-bird.think.com>
In article <············@life.ai.mit.edu> ···@idiap.ch writes:
>In article <··········@festival.ed.ac.uk>, ····@festival.ed.ac.uk (J W Dalton) writes:
>|> Humm. I see your point, but Dylan doesn't just resemble Lisp: right
>|> now, it's a dialect _of_ Lisp.
>If you define "dialect _of_ Lisp" by "having lots of parens", you are
>perhaps right.
There's much more to Lisp than parentheses. I think we've had these kinds
of discussions before, and there's never been a strong concensus of what
defines Lisp, but to me some of the distinguishing features of the Lisp
language are:
* Manifest data types and dynamic type dispatching
* Automatic memory management
* Program text as first-class data (enabling use of the first-order language as
the second-order macro language)
* General-purpose structured data types (e.g. conses)
* Extensive (almost exclusive) use of object references rather than copying
values
Some of these may be tempered or restricted at times (e.g. fixnums are
often passed by value rather than reference, and declarations can obviate
the need for some manifest type information and runtime dispatching). But
a language that has most of these seems "Lispish". It seems even more so
if it also has lots of parentheses. :-)
>But, as far as I can tell, Dylan deviates pretty strongly in some areas
>from existing Lisps, and more changes and additions in areas such as
>static type checking are possible. To me, that makes it a different
>language from Lisp, despite its use of sexprs
To me, the difference between Dylan and Common Lisp is not much more than
the difference between Scheme and Common Lisp.
>To put it differently, the differences between Dylan and CL/CLOS,
>Eulisp, or Scheme seem no smaller than the differences between Oberon
>and Ada, which are called "different languages". And in the same way,
>I would consider Lisp 1.5, MACLISP, and CommonLisp "different
>languages".
And I consider them dialects of Lisp. The designers of these languages
clearly thought that there was enough in common between them that they
should share the "Lisp" in their names. And whatever it is that they all
share, I also see in Scheme and Dylan.
If you don't consider these to be dialects, then how similar do two
languages have to be before you consider them dialects, but not so similar
that they're the same language? Do you consider Symbolics CL, Lucid CL,
CMU CL, and Allegro CL to be different dialects? I think most of us would
consider them to be differing implementation of the CL dialect of Lisp.
To put it another way, an experienced CL or Scheme programmer would have
relatively little trouble transfering his knowledge to Dylan, but going
from Lisp to C requires a significant paradigm shift.
--
Barry Margolin
System Manager, Thinking Machines Corp.
······@think.com {uunet,harvard}!think!barmar
From: Lawrence G. Mayka
Subject: Re: Dialects vs languages (was Re: Future of CMU Common Lisp)
Date:
Message-ID: <LGM.93Aug4205339@hermit.flw.att.com>
In article <············@early-bird.think.com> ······@think.com (Barry Margolin) writes:
In article <············@life.ai.mit.edu> ···@idiap.ch writes:
>To put it differently, the differences between Dylan and CL/CLOS,
>Eulisp, or Scheme seem no smaller than the differences between Oberon
>and Ada, which are called "different languages". And in the same way,
>I would consider Lisp 1.5, MACLISP, and CommonLisp "different
>languages".
And I consider them dialects of Lisp. The designers of these languages
clearly thought that there was enough in common between them that they
should share the "Lisp" in their names. And whatever it is that they all
share, I also see in Scheme and Dylan.
If you don't consider these to be dialects, then how similar do two
languages have to be before you consider them dialects, but not so similar
that they're the same language? Do you consider Symbolics CL, Lucid CL,
CMU CL, and Allegro CL to be different dialects? I think most of us would
consider them to be differing implementation of the CL dialect of Lisp.
I am amazed by your naivete.
In the parallel realms of human languages and programming languages,
the "dialect" tag is highly derogatory, indeed libelous. In both
realms, it implies:
- Not standardized
- Avoided and scorned by the well-educated
- No recognized body of literature
- No distinct cultural heritage
Our adversaries, including highly prestigious C++ programmer/advocates
with best-selling books to their name, have been very successful in
contemptuously referring obliquely but repeatedly in print to
draft-ANSI Common Lisp as a "dialect of Lisp"--to our considerable
detriment. For heaven's sake, we shouldn't agree with them!
We should insist that draft-ANSI Common Lisp is no more a "dialect of
Lisp" than Ukrainian is a "dialect of Russian" or Dutch is a "dialect
of German."
A reasonable term for related =languages= is a "language family."
Lisp is a language family just as East Slavic and West Germanic are
language families.
Lawrence G. Mayka
AT&T Bell Laboratories
···@iexist.att.com
Standard disclaimer.
--
Lawrence G. Mayka
AT&T Bell Laboratories
···@iexist.att.com
Standard disclaimer.
In article <················@hermit.flw.att.com> ···@hermit.flw.att.com (Lawrence G. Mayka) writes:
|I am amazed by your naivete.
And I by your - what? Is this ignorance, arrogance, or just a case of
cynicism worse even than my own?
|In the parallel realms of human languages and programming languages,
|the "dialect" tag is highly derogatory, indeed libelous. In both
|realms, it implies:
|
|- Not standardized
|
|- Avoided and scorned by the well-educated
|
|- No recognized body of literature
|
|- No distinct cultural heritage
No serious linguist, anthropolgist or programming language designer
would accept this for a moment. Is not the "standard" dialect of a
language a dialect? Are all dialects not regulated - standardised - in
just the same way, namely by the requirement of successful communication
(yes, a few languages - notably french and finnish - labour under
government regulation. But compare you the number of times the
government knuckles under on some point of usage to the number that the
populace does...)? If "dialects" (in your derogatory sense) can have no
recognised body of literature, how is it that my shelves are full of
works by americans, africans, irishmen, scots, danes and germans, each
one of whom writes in a dialect quite clearly distinct from mine, the
queen's, and that which was once promulgated by the BBC?
And finally, *how* can you ever assert that the geordies, the cockneys,
the scots, the welsh, jamaicans, south africans, new zealanders, texans
and - I'm sure I'm missing a few here - "lack distinct cultural
heritage"? Your line of argument shows itself to be ... bizarre.
|Our adversaries, including highly prestigious C++ programmer/advocates
|with best-selling books to their name, have been very successful in
|contemptuously referring obliquely but repeatedly in print to
|draft-ANSI Common Lisp as a "dialect of Lisp"--to our considerable
|detriment. For heaven's sake, we shouldn't agree with them!
|
|We should insist that draft-ANSI Common Lisp is no more a "dialect of
|Lisp" than Ukrainian is a "dialect of Russian" or Dutch is a "dialect
|of German."
Points like this are (sadly enough) indeed debated in the natural
language domain. Admittedly this shows a certain failure to understand
the mechanisms of evolution (it's impressive that something so widely
taught should be so little understood, but that's a hobbyhorse of a
divergent hue); please note that indeed speakers can be found of whom it
cannot be decided whether their language be german or dutch - it has
rightly been said that a language is a dialect with a navy, but today it
is the line on the map and not the navy per se that makes all the
difference.
What is most interesting, though, is that the argument is usually in the
opposite direction from your implied line: people will say "scots is no
mere dialect of english, it is a language in its own right!". But if we
were to say "draft-ANSI Common Lisp is no dialect of Lisp, it is a
language in its own right", would we not be playing straight into the
hands of those who would imply that this standard will be no standard at
all, but a fabrication from whole cloth?
Not that any two dialects of C++ are especially compatible....
|A reasonable term for related =languages= is a "language family."
|Lisp is a language family just as East Slavic and West Germanic are
|language families.
The question is as to how far back one needs to go to find a common
ancestor. If they are mutually comprehensible _today_, perhaps "dialect"
is a good word. Otherwise, terms like "family" and "group" are
applicable. Could a programmer debug a simple programme in the other
language without the manual handy (this corresponds to buying fish
without a dictionary, I think)? Yes? Then these are dialects.
I rest my case.
----------------------------------------------------------------------
stephen p spackman +49 681 302 5288(o) 5282(sec) ·······@acm.org
dfki +1.24 / stuhlsatzenhausweg 3 / 66123 saarbruecken / germany
----------------------------------------------------------------------
From: Lawrence G. Mayka
Subject: Re: Dialects vs languages (was Re: Future of CMU Common Lisp)
Date:
Message-ID: <LGM.93Aug9170934@hermit.flw.att.com>
In article <·····················@disco-sol.dfki.uni-sb.de> ········@disco-sol.dfki.uni-sb.de (Stephen P Spackman) writes:
In article <················@hermit.flw.att.com> ···@hermit.flw.att.com (Lawrence G. Mayka) writes:
|I am amazed by your naivete.
And I by your - what? Is this ignorance, arrogance, or just a case of
cynicism worse even than my own?
Perhaps paranoia, justified by experience.
|In the parallel realms of human languages and programming languages,
|the "dialect" tag is highly derogatory, indeed libelous. In both
|realms, it implies:
|
|- Not standardized
|
|- Avoided and scorned by the well-educated
|
|- No recognized body of literature
|
|- No distinct cultural heritage
No serious linguist, anthropolgist or programming language designer
would accept this for a moment. Is not the "standard" dialect of a
language a dialect? Are all dialects not regulated - standardised - in
just the same way, namely by the requirement of successful communication
(yes, a few languages - notably french and finnish - labour under
government regulation. But compare you the number of times the
government knuckles under on some point of usage to the number that the
populace does...)? If "dialects" (in your derogatory sense) can have no
recognised body of literature, how is it that my shelves are full of
works by americans, africans, irishmen, scots, danes and germans, each
one of whom writes in a dialect quite clearly distinct from mine, the
queen's, and that which was once promulgated by the BBC?
And finally, *how* can you ever assert that the geordies, the cockneys,
the scots, the welsh, jamaicans, south africans, new zealanders, texans
and - I'm sure I'm missing a few here - "lack distinct cultural
heritage"? Your line of argument shows itself to be ... bizarre.
My Webster's dictionary lists several definitions of "dialect," almost
all of them stating or implying nonconformance to a standard. English
writings by authors of different nationalities and cultures certainly
show some variation in usage, but I would not denigrate such
literature by referring to it as "written in the British (Irish,
Scottish, etc.) dialect of English." If the author actually intended
to write in a separate language--one related to but different from
standard English--I would then say that he/she wrote in "the so-called
Texan (etc.) language." That is the point: Well-educated authors
rarely =intend= to write in a "dialect" (except occasionally with
respect to quoted speech).
|Our adversaries, including highly prestigious C++ programmer/advocates
|with best-selling books to their name, have been very successful in
|contemptuously referring obliquely but repeatedly in print to
|draft-ANSI Common Lisp as a "dialect of Lisp"--to our considerable
|detriment. For heaven's sake, we shouldn't agree with them!
|
|We should insist that draft-ANSI Common Lisp is no more a "dialect of
|Lisp" than Ukrainian is a "dialect of Russian" or Dutch is a "dialect
|of German."
Points like this are (sadly enough) indeed debated in the natural
language domain. Admittedly this shows a certain failure to understand
the mechanisms of evolution (it's impressive that something so widely
taught should be so little understood, but that's a hobbyhorse of a
divergent hue); please note that indeed speakers can be found of whom it
cannot be decided whether their language be german or dutch - it has
rightly been said that a language is a dialect with a navy, but today it
is the line on the map and not the navy per se that makes all the
difference.
What is most interesting, though, is that the argument is usually in the
opposite direction from your implied line: people will say "scots is no
mere dialect of english, it is a language in its own right!". But if we
were to say "draft-ANSI Common Lisp is no dialect of Lisp, it is a
language in its own right", would we not be playing straight into the
hands of those who would imply that this standard will be no standard at
all, but a fabrication from whole cloth?
I would rather say that "Common Lisp is a language that standardizes a
number of earlier dialects in the Lisp family."
Not that any two dialects of C++ are especially compatible....
If we were to play the same game as that of the C++ advocates I
described, we could refer to various C++ constructs as "the usage in
some dialects of C."
|A reasonable term for related =languages= is a "language family."
|Lisp is a language family just as East Slavic and West Germanic are
|language families.
The question is as to how far back one needs to go to find a common
ancestor. If they are mutually comprehensible _today_, perhaps "dialect"
is a good word. Otherwise, terms like "family" and "group" are
applicable. Could a programmer debug a simple programme in the other
language without the manual handy (this corresponds to buying fish
without a dictionary, I think)? Yes? Then these are dialects.
As I tried to explain earlier, though, this is not the criterion used
by others, in either the human or programming language realms.
--
Lawrence G. Mayka
AT&T Bell Laboratories
···@iexist.att.com
Standard disclaimer.
In article <················@hermit.flw.att.com> ···@hermit.flw.att.com (Lawrence G.
Mayka) writes:
> In article <·····················@disco-sol.dfki.uni-sb.de>
········@disco-sol.dfki.uni-sb.de (Stephen P Spackman) writes:
>
> In article <················@hermit.flw.att.com> ···@hermit.flw.att.com
(Lawrence G. Mayka) writes:
>
> |I am amazed by your naivete.
>
> And I by your - what? Is this ignorance, arrogance, or just a case of
> cynicism worse even than my own?
>
> Perhaps paranoia, justified by experience.
>
> |In the parallel realms of human languages and programming languages,
> |the "dialect" tag is highly derogatory, indeed libelous. In both
> |realms, it implies:
> |
> |- Not standardized
> |
> |- Avoided and scorned by the well-educated
> |
> |- No recognized body of literature
> |
> |- No distinct cultural heritage
>
> No serious linguist, anthropolgist or programming language designer
> would accept this for a moment. Is not the "standard" dialect of a
> language a dialect? Are all dialects not regulated - standardised - in
> just the same way, namely by the requirement of successful communication
> (yes, a few languages - notably french and finnish - labour under
> government regulation. But compare you the number of times the
> government knuckles under on some point of usage to the number that the
> populace does...)? If "dialects" (in your derogatory sense) can have no
> recognised body of literature, how is it that my shelves are full of
> works by americans, africans, irishmen, scots, danes and germans, each
> one of whom writes in a dialect quite clearly distinct from mine, the
> queen's, and that which was once promulgated by the BBC?
>
> And finally, *how* can you ever assert that the geordies, the cockneys,
> the scots, the welsh, jamaicans, south africans, new zealanders, texans
> and - I'm sure I'm missing a few here - "lack distinct cultural
> heritage"? Your line of argument shows itself to be ... bizarre.
>
> My Webster's dictionary lists several definitions of "dialect," almost
> all of them stating or implying nonconformance to a standard. English
> writings by authors of different nationalities and cultures certainly
> show some variation in usage, but I would not denigrate such
> literature by referring to it as "written in the British (Irish,
> Scottish, etc.) dialect of English." If the author actually intended
> to write in a separate language--one related to but different from
> standard English--I would then say that he/she wrote in "the so-called
> Texan (etc.) language." That is the point: Well-educated authors
> rarely =intend= to write in a "dialect" (except occasionally with
> respect to quoted speech).
Well, we're far gone from the subject of lisp here, but this line of reasoning is
too absurd to let pass ... pasted directly from my digital websters (i hope this
constitutes fair use :-)
____
1a: a regional variety of language distinguished by features of vocabulary,
grammar, and pronunciation from other regional varieties and constituting together
with them a single language of which no one variety is construed as standard (the
Doric dialect of ancient Greek)
[Hmm. nothing derogatory here. No "standard" reference language mentioned.]
b: one of two or more cognate languages - French and Italian are Romance dialects
[still nothing derogatory]
c: a regional variety of a language usu. transmitted orally and differing
distinctively from the standard language - the Lancashire dialect of English
[ aha, reference to a standard language, and an example - but still no indication
as too cultural inferiority]
d: a variety of a language used by the members of an occupational group - the
dialect of the atomic physicist
[still neutral]
e: a variety of language whose identity is fixed by a factor other than geography
(as social class or educational level of its habitual users) - spoke a rough peasant
dialect
[ aha ! suggestions of classism, potential for insult here...]
_____
You're reading of 'dialect' as a derogatory seems to be a personal quirk of your
own, citing Websters to support this viewpoint looks pretty thin from where i sit.
Further, let me add that some of my best friends are Texans, and your suggestion
that they are speaking some language other than English I view as being *highly*
insulting. Of course as a speaker of the obscure "cheesehead" dialect of english my
views may not be representitive :-)
Lee Schumacher
In article <······················@schaefer.math.wisc.edu> ········@math.wisc.edu writes:
You're reading of 'dialect' as a derogatory seems to be a personal quirk of your
own, citing Websters to support this viewpoint looks pretty thin from where i sit.
If you were a Ukrainian who was told by a Russian imperialist that
"Ukrainian is just a dialect of Russian" (with the implication that
Russia has the right to take over Ukraine), you might feel
differently.
Further, let me add that some of my best friends are Texans, and your suggestion
that they are speaking some language other than English I view as being *highly*
insulting. Of course as a speaker of the obscure "cheesehead" dialect of english my
views may not be representitive :-)
This =new= misunderstanding only underscores the strong emotions
attached to the whole issue.
--
Lawrence G. Mayka
AT&T Bell Laboratories
···@iexist.att.com
Standard disclaimer.
From: Thomas M. Breuel
Subject: Re: Dialects vs languages (was Re: Future of CMU Common Lisp)
Date:
Message-ID: <TMB.93Aug5192026@arolla.idiap.ch>
In article <············@early-bird.think.com> ······@think.com (Barry Margolin) writes:
|>If you define "dialect _of_ Lisp" by "having lots of parens", you are
|>perhaps right.
|
|There's much more to Lisp than parentheses.
Do I have to put smileys everywhere?
|I think we've had these kinds
|of discussions before, and there's never been a strong concensus of what
|defines Lisp, but to me some of the distinguishing features of the Lisp
|language are:
|
|* Manifest data types and dynamic type dispatching
|* Automatic memory management
|* Program text as first-class data (enabling use of the first-order language as
| the second-order macro language)
|* General-purpose structured data types (e.g. conses)
|* Extensive (almost exclusive) use of object references rather than copying
| values
|
|Some of these may be tempered or restricted at times (e.g. fixnums are
|often passed by value rather than reference, and declarations can obviate
|the need for some manifest type information and runtime dispatching). But
|a language that has most of these seems "Lispish". It seems even more so
|if it also has lots of parentheses. :-)
Thinking about a number of languages that I'm familiar with, these
conditions, neither individually, or in groups seem to be either
necessary or sufficient for a language to be called "Lisp".
In any case, when talking about terminology, we should look at what
the terminology is going to be used for. In this case, Jeff (I think)
stated first that "Dylan was Lisp" and then turned around and asked
why Apple didn't just use Eulisp. I think that answers the question:
Dylan isn't Lisp because Apple doesn't want it to be; they want to
have the freedom to deviate in arbitrary ways from Lisp tradition, and
they don't want to be told by opponents of Lisp "Dylan is bad because
it is Lisp", and they don't want to be told by proponents of Lisp that
"Dylan is bad because it isn't like Lisp in some areas".
|To put it another way, an experienced CL or Scheme programmer would have
|relatively little trouble transfering his knowledge to Dylan, but going
|from Lisp to C requires a significant paradigm shift.
I think Dylan is much closer to Smalltalk, Eiffel, or CLOS than to
CommonLisp or Scheme, at least as far as "paradigms" go.
As far as I can tell, Dylan is going to have a different syntax, a
different "vocabulary" (function names), a different type system
(dynamic and static), a different paradigm (object-oriented vs.
functional, higher-order, or procedural), and a different user
community from Scheme or Lisp. I think its relationship to Lisp is
well covered by its name "a Dynamic (object-oriented) Language"--it's
no more and no less.
Thomas.
From: J W Dalton
Subject: Re: Dialects vs languages (was Re: Future of CMU Common Lisp)
Date:
Message-ID: <CBAv3p.11s@festival.ed.ac.uk>
···@arolla.idiap.ch (Thomas M. Breuel) writes:
>Thinking about a number of languages that I'm familiar with, these
>conditions, neither individually, or in groups seem to be either
>necessary or sufficient for a language to be called "Lisp".
Who says there have to be such conditions? I recall Wittgenstein's
remarks on "game" and think in terms of "family resemblance". In
any case, I don't think it's worth spending much time arguing over
what, if any, conditions are necessary or sufficient. I will say
more about this in another message.
>In any case, when talking about terminology, we should look at what
>the terminology is going to be used for. In this case, Jeff (I think)
>stated first that "Dylan was Lisp" and then turned around and asked
>why Apple didn't just use Eulisp.
I don't think I've ever said Dylan was Lisp. Was _a_ Lisp, or a
_variety of_ Lisp, or a dialect of Lisp, or a member of the Lisp
family / group of languages, yes; but not Dylan _is_ Lisp.
I don't think I have never asked why Apple didn't just use EuLisp.
You must have confused me with someone else.
> I think that answers the question:
>Dylan isn't Lisp because Apple doesn't want it to be;
I see. So this thing I'm typing on isn't a keyboard if I don't
want it to be?
>they want to have the freedom to deviate in arbitrary ways from
>Lisp tradition,
They can deviate as much as they'd like (though a worse language
may result); but until they do they have a variety of Lisp.
> and
>they don't want to be told by opponents of Lisp "Dylan is bad because
>it is Lisp", and they don't want to be told by proponents of Lisp that
>"Dylan is bad because it isn't like Lisp in some areas".
I don't think any Lisp advocate thinks a language is bad merely
because it isn't like Lisp. The "some areas" aren't going to
be left blank. We might think a language is bad because it
lacks certain properties, where Lisp does not lack them, but
we don't think those properties are good because Lisp has them;
instead we think Lisp is good because it has those properties.
On the other hand, people often do conclude that something is bad
because it's Lisp. I can understand Apple wanting to avoid this
prejudiced reaction.
Moreover, I very seldom say languages are bad. I typically say
there's more than one good kind of languges and that if I prefer
one kind, other people can reasonably prefer others.
>|To put it another way, an experienced CL or Scheme programmer would have
>|relatively little trouble transfering his knowledge to Dylan, but going
>|from Lisp to C requires a significant paradigm shift.
>I think Dylan is much closer to Smalltalk, Eiffel, or CLOS than to
>CommonLisp or Scheme, at least as far as "paradigms" go.
CLOS is part of Common Lisp, not a separate language, even though
people sometimes speak as if it were otherwise. But in case someone
insists on disputing this, TELOS in EuLisp and the explicit OOP
support in ISLisp (the possible ISO standard) are not late additions
(as CLOS was).
You will note that when I list the Lisps Dylan is similar to, it's
EuLisp, ISLisp, and Scheme + a subset of CLOS.
>As far as I can tell, Dylan is going to have a different syntax, a
>different "vocabulary" (function names), a different type system
>(dynamic and static), a different paradigm (object-oriented vs.
>functional, higher-order, or procedural), and a different user
>community from Scheme or Lisp. I think its relationship to Lisp is
>well covered by its name "a Dynamic (object-oriented) Language"--it's
>no more and no less.
You persist in speaking of Lisp as if it were some particular language,
presumably one predating object-oritented Lisps such as EuLisp and
Dylan. But this is wrong. If Lisp is any one thing, it's a family
of languages; and Lisp has always evolved to encompass new ways of
doing things such as object-oriented programming. People who dislike
Lisp may wish this weren't so, and often talk as if it weren't,
but they are wrong, and wanting something doesn't make it so.
-- jd