From: Adrian B.
Subject: What’s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <7ed8f64d.0209011823.6c7ceeca@posting.google.com>
As a casual user of scheme and reader of the related newsgroups, I've
noticed that there's always so much debate, confusion, and critiques
of the scheme macro system.

Is there something fundamentally broken with Scheme macros?  Are they
badly designed?  Or is all this discussion simply due to confusion?
 
Why the constant debate?

If the macro system is badly designed, are there any serious attempts
to improve it?   Or is it cast in stone forever more due to appearing
in the R5RS?  I've seen one alternative, syntax-case, but is it a
serious alternative for the standard, or is it just a slightly better
version of syntax rules?

Finally, is there a Scehme macros FAQ?  I must admit I'm quite
confused by all this debate.

Thanks

From: Adrian B.
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <7ed8f64d.0209020319.1d5fb809@posting.google.com>
····@ma.ccom (Takehiko Abe) wrote in message news:<·····················@solg4.keke.org>...
> In article <····························@posting.google.com>, ····@swirve.com (Adrian B.) wrote:
> 
> > Is there something fundamentally broken with Scheme macros? 
> 
> Please do not cross post to comp.lang.lisp.

And why exactly not?  Lispers are strong users of macros and may have
some good experience on the design and use of macro systems.  Its not
like the question was crossposted to a C++ or Java newsgroup.
From: Tim Bradshaw
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <ey3k7m4oc6b.fsf@cley.com>
* Adrian B wrote:

> And why exactly not?

Years of bitter experience indicate that no good comes of these
discussions.  Look at google.

c.l.l seems to be reverting to type alas - we have our first genuine
semi-literate but arrogant halfwit in ages (ilias), and now we are
about to have a c.l.l / c.l.s flame war.

--tim
From: Kenny Tilton
Subject: Ilias alas? Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3D739165.90105@nyc.rr.com>
I for one do not despair over the arrival of ilias on our shore, for he 
has many redeeming qualties. He loves Lisp (or at least professes it). 
He does not go back-and-forth ad nauseum over a personal dispute. His 
articles are mostly technical. Best of all, his articles are short.

As for his refusal to learn, well, that is a self-limiting quality. One 
by one all who would offer guidance discover the futility of the effort 
and thereafter lurk him for amusement value.

But what is the sound of one Ilias corresponding?

:)

kenny

Tim Bradshaw wrote:

> * Adrian B wrote:
> 
> 
>>And why exactly not?
>>
> 
> Years of bitter experience indicate that no good comes of these
> discussions.  Look at google.
> 
> c.l.l seems to be reverting to type alas - we have our first genuine
> semi-literate but arrogant halfwit in ages (ilias), and now we are
> about to have a c.l.l / c.l.s flame war.
> 
> --tim
> 
> 
From: Thien-Thi Nguyen
Subject: Re: Ilias alas? Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <kk9d6rwyzcl.fsf@glug.org>
Kenny Tilton <·······@nyc.rr.com> writes:

> But what is the sound of one Ilias corresponding?

people are attached to their own learning model and sometimes can't make light
of others', but all crystalographers know the value of refraction and internal
reflection.  like all gems, lisp is amenable to such methods.  lisp experts
can drink diamonds and thus sometimes forget its liquid properties are due to
their expertise and practice, which is applicable mostly on a personal basis.

i'm glad ilias did not post a 2-line sed (or other text transform) script!

thi,
just another H for M-x blackbox-usenet...
From: ilias
Subject: Re: Ilias alas? Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3D73EBBC.1060800@pontos.net>
Thien-Thi Nguyen wrote:
> Kenny Tilton <·······@nyc.rr.com> writes:
> 
> 
>>But what is the sound of one Ilias corresponding?
> 
> 
> people are attached to their own learning model and sometimes can't make light
> of others', but all crystalographers know the value of refraction and internal
> reflection.  like all gems, lisp is amenable to such methods.  lisp experts
> can drink diamonds and thus sometimes forget its liquid properties are due to
> their expertise and practice, which is applicable mostly on a personal basis.
> 
> i'm glad ilias did not post a 2-line sed (or other text transform) script!

can someone please be so kindly to transform this to simple English?

(i'm just assimilating the LISP-reader, and i'm unable to switch 
context. but i'd like to know.)
From: Software Scavenger
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <a6789134.0209021413.9956022@posting.google.com>
····@swirve.com (Adrian B.) wrote in message news:<····························@posting.google.com>...

> And why exactly not?  Lispers are strong users of macros and may have
> some good experience on the design and use of macro systems.  Its not
> like the question was crossposted to a C++ or Java newsgroup.

I agree that it's interesting to discuss differences in macros in
different languages.  It seems easy enough for those who aren't
interested to just skip this whole thread.  Especially with such an
easily plonkable thread name that won't even cause any hesitation.

As an example to compare Common Lisp macros with Scheme macros,
consider a macro named DO-COMBINATIONS, which works as follows:

(do-combinations (x1 x2 x3) '(1 2 3 4 5)
   (print (list x1 x2 x3)))
===>
(1 2 3) 
(1 2 4) 
(1 2 5) 
(1 3 4) 
(1 3 5) 
(1 4 5) 
(2 3 4) 
(2 3 5) 
(2 4 5) 
(3 4 5)

(do-combinations (a b)
   '(heart spade diamond club)
      (print (list a b)))
===>
(HEART SPADE) 
(HEART DIAMOND) 
(HEART CLUB) 
(SPADE DIAMOND) 
(SPADE CLUB) 
(DIAMOND CLUB)

(defun combinations-of-3 (list)
   (let (result)
      (do-combinations (a b c) list
         (push (list a b c) result))
      (reverse result)))

(combinations-of-3 '(1 2 3 4 5))
===>
((1 2 3) (1 2 4) (1 2 5) (1 3 4) (1 3 5) (1 4 5) (2 3 4) (2 3 5) (2 4
5) (3 4 5))

The above three examples show three ways to use the example
DO-COMBINATIONS macro.

I don't actually know Scheme, but would like to see the implementation
of the above DO-COMBINATIONS macro in it.  Then I or someone could
post a CL version of it to compare.  Maybe several people could post
more than one way to implement it, to compare different ways in each
language.  Then we could do the same thing for other kinds of macros. 
Etc., till we get a better idea of the differences.
From: ····@pobox.com
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <7eb8ac3e.0209022005.42e8880b@posting.google.com>
··········@mailandnews.com (Software Scavenger) wrote in message news:<···························@posting.google.com>...
> As an example to compare Common Lisp macros with Scheme macros,
> consider a macro named DO-COMBINATIONS, which works as follows:
> 
> (do-combinations (x1 x2 x3) '(1 2 3 4 5)
>    (print (list x1 x2 x3)))
> I don't actually know Scheme, but would like to see the implementation
> of the above DO-COMBINATIONS macro in it.  

; The do-combinations macro
; (do-combinations (x1 x2 x3) (1 2 3 4 5)
;   (print (list x1 x2 x3)))
; will print all combinations of three elements out of (1 2 3 4 5)

(define-syntax do-combinations
  (syntax-rules ()
    ((_ (binder ...) elements body)
     (for-each
      (lambda (x) (apply (lambda (binder ...) body) x))
      (subsets (length (quote (binder ...)))
			   (quote elements))))))


; create a list of all subsets from 'elements' of length 'how-many'
; The resulting list will have the size of
; (choose how-many (length elements))
; Recurrence relation:
; (subsets how-many (cons el elems)) =
; (append (subsets how-many elems)
;         (map add-el (subsets (-- how-many) elems)))
; (subsets 0 elems) = '()
; (subsets 1 elems) = (map list elems) ; singleton subsets
; (subsets n elems) = '() whenever n > (length elems)
;
; The code is optimized for clarity rather than for speed
; See the thread
; http://groups.google.com/groups?threadm=7eb8ac3e.0201120056.3fc231c8%40posting.google.com
; for the fastest implementation of this function.

(define (subsets how-many elements)
  (cond
   ((zero? how-many) '())
   ((null? elements) '())
   ((= 1 how-many) (map list elements))
   ((= how-many (length elements)) (list elements))
   ((> how-many (length elements)) '())
   (else
    (append (subsets how-many (cdr elements))
	    (map (lambda (x) (cons (car elements) x))
		 (subsets (- how-many 1) (cdr elements)))))))

; Test cases
(pp (subsets 0 '(heart spade diamond club)))
(pp (subsets 1 '(heart spade diamond club)))
(pp (subsets 2 '(heart spade diamond club)))
(pp (subsets 3 '(heart spade diamond club)))
(pp (subsets 4 '(heart spade diamond club)))
(pp (subsets 5 '(heart spade diamond club)))

; Examples
(do-combinations (a b)
   (heart spade diamond club)
   (begin
     (display (list a b))
     (newline)))
===>
(diamond club)
(spade diamond)
(spade club)
(heart spade)
(heart diamond)
(heart club)

(do-combinations (x1 x2 x3)
   (1 2 3 4 5)
   (begin
     (display (list x1 x2 x3))
     (newline)))
===>
(3 4 5)
(2 4 5)
(2 3 4)
(2 3 5)
(1 4 5)
(1 3 4)
(1 3 5)
(1 2 3)
(1 2 4)
(1 2 5)
From: Software Scavenger
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <a6789134.0209041419.cfe3b56@posting.google.com>
I've started a new thread for this topic.  The subject line is "Macros
in Common Lisp, Scheme, and other languages".
From: Al Petrofsky
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <871y8b7jyx.fsf@radish.petrofsky.org>
[Followups redirected to comp.lang.scheme]

··········@mailandnews.com (Software Scavenger) writes:
> ····@swirve.com (Adrian B.) wrote:
> 
> > And why exactly not?  Lispers are strong users of macros and may have
> > some good experience on the design and use of macro systems.

> I agree that it's interesting to discuss differences in macros in
> different languages.

There are certainly some readers of comp.lang.lisp who are interested
in comparing scheme and cl macros, but those people should all be
reading comp.lang.scheme as well.  Scheme is not a recent offshoot of
lisp.  Lisp followers have had plenty of time to consider whether or
not they are interested in the directions scheme has taken, and those
who are interested should look in comp.lang.scheme.

> As an example to compare Common Lisp macros with Scheme macros,
> consider a macro named DO-COMBINATIONS, which works as follows:

> (do-combinations (a b)
>    '(heart spade diamond club)
>       (print (list a b)))
> ===>
> (HEART SPADE) 
> (HEART DIAMOND) 
> (HEART CLUB) 
> (SPADE DIAMOND) 
> (SPADE CLUB) 
> (DIAMOND CLUB)

> I don't actually know Scheme, but would like to see the implementation
> of the above DO-COMBINATIONS macro in it.

That example happens to be well-suited to syntax-rules, scheme's
pattern language for macros:

  (define-syntax do-combinations
    (syntax-rules ()
      ((do-combinations () list-expr expr)
       expr)
      ((do-combinations (var . rest-of-vars) list-expr expr)
       (let do-combos-using-all-vars ((vals list-expr))
         (if (not (null? vals))
             (let ((var (car vals))
                   (rest-of-vals (cdr vals)))
               (do-combinations rest-of-vars rest-of-vals expr)
               (do-combos-using-all-vars rest-of-vals)))))))

An alternative approach would be to keep the macro as simple as
possible and let a procedure do all the work:

  (define-syntax do-combinations
    (syntax-rules ()
      ((do-combinations vars list-expr expr)
       (do-combos-proc 'vars list-expr (lambda vars expr)))))

  (define (do-combos-proc vars vals proc)
    (cond ((null? vars) (proc))
          ((not (null? vals))
           (let ((first-val (car vals))
                 (rest-of-vals (cdr vals)))
             (do-combos-proc (cdr vars) rest-of-vals
                             (lambda args (apply proc first-val args)))
             (do-combos-proc vars rest-of-vals proc)))))

-al
From: Marco Antoniotti
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <y6celcb88oy.fsf@octagon.mrl.nyu.edu>
Al Petrofsky <··@petrofsky.org> writes:

> [Followups redirected to comp.lang.scheme]
> 
> ··········@mailandnews.com (Software Scavenger) writes:
> > ····@swirve.com (Adrian B.) wrote:
> > 
> > > And why exactly not?  Lispers are strong users of macros and may have
> > > some good experience on the design and use of macro systems.
> 
> > I agree that it's interesting to discuss differences in macros in
> > different languages.
> 
> There are certainly some readers of comp.lang.lisp who are interested
> in comparing scheme and cl macros, but those people should all be
> reading comp.lang.scheme as well.  Scheme is not a recent offshoot of
> lisp.  Lisp followers have had plenty of time to consider whether or
> not they are interested in the directions scheme has taken, and those
> who are interested should look in comp.lang.scheme.

New directions in Scheme?  Like "let's re-implement the CLHS"?

Sorry,  I am a sucker.  I cannot resist these flame baits :)

Fell frce to send replies to /dev/null :)

Cheers

-- 
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group        tel. +1 - 212 - 998 3488
715 Broadway 10th Floor                 fax  +1 - 212 - 995 4122
New York, NY 10003, USA                 http://bioinformatics.cat.nyu.edu
                    "Hello New York! We'll do what we can!"
                           Bill Murray in `Ghostbusters'.
From: Christopher Browne
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <al3j0f$1m2j1i$1@ID-125932.news.dfncis.de>
The world rejoiced as "Perry E. Metzger" <·····@piermont.com> wrote:
> Feuer <·····@his.com> writes:
>> > In fact, it is not an offshoot at all.  It is a lisp.  May I
>> > assume that you are in the "c.l.l is for CL only" brigade?
>> 
>> Is it really a Lisp?  Does it really retain anything significant from
>> Lisp other than parentheses and call-by-value semantics?
>
> It is hard to see why scheme would not be considered a lisp but emacs
> lisp or interlisp or the lisp in the sawfish window manager or
> whatever would not be. It is certainly not even an unusual dialect --
> lisp 1 also had the unified function/variable namespace, the lexical
> scoping is now the norm rather than unusual, etc. What makes scheme
> the least bit unlispy, in fact?

The fact that such questions lead to flame wars and division, along
quite clearly discernable lines that form on both sides.

There may be things worth discussing about the comparative
similarities and differences between Common Lisp and Scheme; there
might conceivably be some new things worth discussing, even concerning
the respective macro systems.

Unfortunately, past history shows that such discussions generally lead
to flames and acrimony.  

The only way that you should _consider_ such a discussion is if you
are:
 a) Well aware of the possibilities for flames, and
 b) Really quite sure, based on reviewing past discussions, that you
    aren't merely rehashing ancient history.

The fact that you aren't aware of the flameworthiness of the
discussion suggests that it is _highly_ unlikely that you have
reviewed past discussions, and that there is little reason to expect
the discussion to lead anywhere constructive.
-- 
(reverse (concatenate 'string ·············@" "enworbbc"))
http://www.ntlug.org/~cbbrowne/multiplexor.html
Signs of  a Klingon Programmer - 2.  "Specifications are  for the weak
and timid!"
From: Perry E. Metzger
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <87fzwq79dz.fsf@snark.piermont.com>
Christopher Browne <········@acm.org> writes:
> > It is hard to see why scheme would not be considered a lisp but emacs
> > lisp or interlisp or the lisp in the sawfish window manager or
> > whatever would not be. It is certainly not even an unusual dialect --
> > lisp 1 also had the unified function/variable namespace, the lexical
> > scoping is now the norm rather than unusual, etc. What makes scheme
> > the least bit unlispy, in fact?
> 
> The fact that such questions lead to flame wars and division, along
> quite clearly discernable lines that form on both sides.

I find something interesting about both comp.lang.lisp and
comp.lang.scheme. In other language newsgroups, like the the ruby
group or what have you, people spend most of their time asking
questions about how to solve particular problems they're having with
their production code using the language or what have you. This tends
to indicate the people involved are largely interested in programming
languages as a way of accomplishing their work. In the c.l.l and
c.l.s, however, we have groups of people who are concerned largely
about linguistic metaissues, that is, about their language as a
religion.

It would appear that, in their ability to confine themselves almost
entirely at the religious layer of the stack, the two communities are
obviously made up of people of nearly the same mentality. Whether this
is a productive mentality or not remains to be seen, but clearly if we
are to ask a question someone else mentioned today and ask if both
Common Lisp and the Scheme communities share an outlook, there is no
question that they do.

Perhaps to insiders there is a tremendous difference in world view
between the two groups, but to outsiders it looks ever so much like
two groups of Christians a few hundred years ago persecuting each
other over doctrinal minutia utterly unimportant to the overall
picture.

> The fact that you aren't aware of the flameworthiness of the
> discussion suggests that it is _highly_ unlikely that you have
> reviewed past discussions, and that there is little reason to expect
> the discussion to lead anywhere constructive.

Perhaps we could have a new discussion about why it is that people
think that religion is more interesting than writing programs.

-- 
Perry E. Metzger		·····@piermont.com
--
"Ask not what your country can force other people to do for you..."
From: Erik Naggum
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3240097371783391@naggum.no>
* "Perry E. Metzger" <·····@piermont.com>
| Perhaps we could have a new discussion about why it is that people
| think that religion is more interesting than writing programs.

  Perhaps people who use Common Lisp and Scheme are already adept at writing
  programs and no longer have the pathetic little problems that so plague the
  newer languages and are interested in politics (not religion) just as people
  who think there is a better way to make a living than wonder how to get a
  job serving burgers get into law and politics and diplomacy and standards.
  That you have no better grasp of the higher levels of society than to call
  it "religion" unfortunately speaks volumes about your own outlook on that
  which transcends petty problems expressing simple algorithms in languages
  younger than their practitioners' computers.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Hrvoje Blazevic
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <al4h8a$afan$1@as201.hinet.hr>
Erik Naggum wrote:
>   Perhaps people who use Common Lisp and Scheme are already adept at writing
>   programs and no longer have the pathetic little problems that so plague the
>   newer languages and are interested in politics (not religion) just as people
>   who think there is a better way to make a living than wonder how to get a
>   job serving burgers get into law and politics and diplomacy and standards.

Umm... That is an interesting view. I would rather say that people who think 
there is a better way to make a living then wonder how to get a job serving 
burgers get into LAW and POLITICS and DIPLOMACY ... are people who are not 
interested in working at all.

As for politics not being religion ... I was "fortunate" enough to have been 
born in a communist run country, and know very well that communism (probably 
most of the other "politics" as well) is very much a religion at its best -- 
cause if you speak up; they will burn you!

Hrvoje
From: Erik Naggum
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3240158074254685@naggum.no>
* Hrvoje Blazevic
| Umm... That is an interesting view. I would rather say that people who think
| there is a better way to make a living then wonder how to get a job serving
| burgers get into LAW and POLITICS and DIPLOMACY ... are people who are not
| interested in working at all.

  What was the significance of the "then" that replaced my "than"?  Presuming
  it has meaning, I am unable to understand what you mean fully.  I also
  wonder what kind of misguided notions "work" that make up the substance of
  your opinion. May I offer a counter-view that muscle-time is irrelevant and
  the only measure of the amount of work involved is the brain-time needed.

| As for politics not being religion ... I was "fortunate" enough to have been
| born in a communist run country, and know very well that communism (probably
| most of the other "politics" as well) is very much a religion at its best --
| cause if you speak up; they will burn you!

  It is religion that has elements of politics, not the other way around.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Hrvoje Blazevic
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <al5rhg$bmht$1@as201.hinet.hr>
Erik Naggum wrote:
>   What was the significance of the "then" that replaced my "than"?  Presuming
>   it has meaning, I am unable to understand what you mean fully.

Just a typo!

> I also
>   wonder what kind of misguided notions "work" that make up the substance of
>   your opinion. May I offer a counter-view that muscle-time is irrelevant and
>   the only measure of the amount of work involved is the brain-time needed.

I also wonder what led you to believe that I would value selling burgers more 
than the brain-time. Is it because you believe that it is only lawyers, 
politicians and diplomats that do use brain-power?
From: Erik Naggum
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3240164989340408@naggum.no>
* Hrvoje Blazevic
| I also wonder what led you to believe that I would value selling burgers
| more than the brain-time.  Is it because you believe that it is only lawyers,
| politicians and diplomats that do use brain-power?

  No.  I consider this "discussion" to have ended some time ago.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: William D Clinger
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <b84e9a9f.0209041428.35cde03@posting.google.com>
Erik Naggum wrote:
>   Perhaps people who use Common Lisp and Scheme are already adept at writing
>   programs and no longer have the pathetic little problems that so plague the
>   newer languages....

Isn't it pretty to think so?

Ernest Hemingway
From: Erik Naggum
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3240169951896806@naggum.no>
* William D Clinger
| Isn't it pretty to think so?
| 
| Ernest Hemingway

  Nice quotation.  Sometimes, I think people think ugly because they do not
  want pretty to be an option in the real world.  There is a implicit accusation
  in this quotation that if it is pretty, it cannot be true, and if you believe
  the pretty option, you are na�ve.  This attitude is unfortunately very hard
  to combat, since those who believe in such things are amazingly unwilling to
  accept that the world can have pretty parts when their whole life experience
  has been that none of it is.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Perry E. Metzger
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <878z2h5wkw.fsf@snark.piermont.com>
Erik Naggum <····@naggum.no> writes:
> * "Perry E. Metzger" <·····@piermont.com>
> | Perhaps we could have a new discussion about why it is that people
> | think that religion is more interesting than writing programs.
> 
>   Perhaps people who use Common Lisp and Scheme are already adept at
>   writing programs and no longer have the pathetic little problems
>   that so plague the newer languages

Perhaps, but I doubt it. I've been writing software in a variety of
languages for... well, I don't want to think about it but I started on
PDP-8s a long time ago. I still learn new things all the time. Hell, I
even learn whole new paradigms.

What I've noticed in a lot of people in this community (and it really
is just one community) is a lot of arrogance. Being a very arrogant
person at times, I can perhaps recognize the symptom a bit better than
most. I have often made the mistake of thinking I knew more than I
really did, but I've been working pretty hard on keeping my mind open
instead, and it often brings results.

>   and are interested in politics (not religion) just as people
>   who think there is a better way to make a living than wonder how to get a
>   job serving burgers get into law and politics and diplomacy and standards.

I don't know about that. I'd expect that if the community was truly
vibrant we'd be seeing things like equivalents to CPAN and such. It is
not disgraceful to say "I'm writing an application that needs to
interface with database X, anyone have a binding written already" or
what have you. Perhaps the Lisp community is beyond actually having to
use its language day to day but I doubt that.

>   That you have no better grasp of the higher levels of society than to call
>   it "religion" unfortunately speaks volumes about your own outlook on that
>   which transcends petty problems expressing simple algorithms in languages
>   younger than their practitioners' computers.

I've been programming long enough to know that when you're more
concerned about what the right comment character is than about writing
good comments you're not on the level of the important any longer. I
also know that the nuts and bolts of getting work done is the petty
problem of expressing algorithms for execution by machine, not the
discussion of whether the guys who use language-flavor Y are apostates
who must be banned from the church.

Software, unlike theoretical mathematics, is largely about
accomplishing things. That leads people to unfortunate mundane
concerns like finding a module that builds web pages for you or
finding a module that interfaces to Oracle or finding a module that
does statistical analysis for you. When a language's devotees no
longer discuss such pragmatic matters and instead spend all their
energy on religion, it implies they are not writing code.


-- 
Perry E. Metzger		·····@piermont.com
--
"Ask not what your country can force other people to do for you..."
From: Thomas Bushnell, BSG
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <87u1l5csr9.fsf@becket.becket.net>
"Perry E. Metzger" <·····@piermont.com> writes:

> I've been programming long enough to know that when you're more
> concerned about what the right comment character is than about writing
> good comments you're not on the level of the important any longer. I
> also know that the nuts and bolts of getting work done is the petty
> problem of expressing algorithms for execution by machine, not the
> discussion of whether the guys who use language-flavor Y are apostates
> who must be banned from the church.

I think the point is that an important goal for Scheme is to do it
right, dammit, or not at all.  This would be horrible if Scheme were
the only choice, but it's not.  Since there are a jillion decent
programming languages with the attitude "who cares what the comment
character is", what's wrong with there being *one* that wants to get
it right, dammit.
From: sv0f
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <none-0409022040190001@129.59.212.53>
In article <··············@snark.piermont.com>, "Perry E. Metzger"
<·····@piermont.com> wrote:

>Perhaps, but I doubt it. I've been writing software in a variety of
>languages for... well, I don't want to think about it but I started on
>PDP-8s a long time ago.

Translation: I'm a veteran programmer.  You're not.

>What I've noticed in a lot of people in this community (and it really
>is just one community) is a lot of arrogance.

T: To me, an old and wise outsider, you guys are arrogant.

>I don't know about that. I'd expect that if the community was truly
>vibrant we'd be seeing things like equivalents to CPAN and such.

T: I like Perl.

>I've been programming long enough to know that when you're more
>concerned about what the right comment character is than about writing
>good comments you're not on the level of the important any longer.

T: You'all listen close cuz, once again, I've been programming
a long time.
From: Erik Naggum
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3240169041516327@naggum.no>
* Perry E. Metzger
| What I've noticed in a lot of people in this community (and it really
| is just one community) is a lot of arrogance.

  Funny, that, I see a lot of humility and respect for the opinions and
  thoughts of those older and wiser than oneself.  There is virtually no
  arrogance among the cognoscenti.  My experience has been one of great
  patience and willingness to help those who wish to learn.  However, when
  some arrogant little snot of a newbie walks into the forum only to spout his
  own arrogant views without listening to anybody else, the /newbie/ will
  think this is arrogance.  If you are an arrogant newbie, people around you
  will generally return the favor.  If you are interested in learning things
  from those who know, the Lisp communities are better than most.  Well, at
  least comp.lang.lisp.  The comp.lang.scheme people have generally been on
  the defensive even when they are not attacked in any way, which I can only
  interpret as symptoms of a very /protective/ crowd.  However, it has never
  been /arrogance/ that I have found even in the hostility towards comments on
  the severely limited Scheme model.

| Being a very arrogant person at times, I can perhaps recognize the symptom a
| bit better than most.

  Perhaps you do much too well?  Arrogant people tend to trigger arrogant
  responses from others, too.  It is a stupid human tendency that is only
  changed by limiting or removing one's own arrogance.

| I have often made the mistake of thinking I knew more than I really did, but
| I've been working pretty hard on keeping my mind open instead, and it often
| brings results.

  Most of the people I have found in the Common Lisp community have been well
  aware, not only of how much they know, but of where it came from and how
  they arrived at the conclusions they hold, so that they can communicate this
  to others when they need to defend their opinions.  Rapid internalization of
  new knowledge and the ability to integrate it with other information make it
  possible to produce intelligent observations and connections.  All of these
  attributes indicate both "intelligent" and "observant" to me, and not the
  dumb and judgmental attributes that correlate with arrogance.

| I don't know about that.  I'd expect that if the community was truly vibrant
| we'd be seeing things like equivalents to CPAN and such.

  I think this shows your arrogance more than anything else.  You are obviously
  not willing to consider the fact that CPAN and the like depend on much more
  than "vibrant communities" to exist.  In particular, they depend on the
  comparative worthlessness of the time spent on making software for their
  languages.  When the time you spend writing software is both worth paying
  for and it is worth paying others for their time, you get a vibrant /market/,
  not a vibrant library of free software.  You also do not solve problems that
  people are willing to solve for free.

| I've been programming long enough to know that when you're more concerned
| about what the right comment character is than about writing good comments
| you're not on the level of the important any longer.

  Just so we have this clear: You think somebody is arguing over the right
  comment character?  Where did this happen? And moreover, why is this so
  important to you that you think it reflects on the whole community?

| I also know that the nuts and bolts of getting work done is the petty
| problem of expressing algorithms for execution by machine, not the
| discussion of whether the guys who use language-flavor Y are apostates who
| must be banned from the church.

  And just so we have this clear, too: This happened where?  And it was
  important to you why?  Perhaps I do not understand your problems, like I do
  not understand the most recent troll in comp.lang.lisp who whines about the
  syntax.  Trolls are not representative of the community in any way.  You
  should know this, I think.

| Software, unlike theoretical mathematics, is largely about accomplishing
| things.  That leads people to unfortunate mundane concerns like finding a
| module that builds web pages for you or finding a module that interfaces to
| Oracle or finding a module that does statistical analysis for you. When a
| language's devotees no longer discuss such pragmatic matters and instead
| spend all their energy on religion, it implies they are not writing code.

  I would like you to think about and enumerate the many other assumptions
  that went into this rather strange conclusion.  The most important (at least
  to refute your conclusion) assumption is why these things are important to
  you.  You are undoutedly right that these things happen, just as we find the
  most brilliant mind sometimes uttering stupid comments unwittingly, but does
  that prove anything?  I do not in any way wish to deny that these things are
  sometimes discussed and are consuming the time of many people when they do,
  but I wonder why you select these events and similarly ignore the events when
  people discuss application-oriented aspects equally consumingly.  I tend to
  think of the former as supporting the notion that people want to think in
  their languages an therefore need syntax that fits their thinking.  It is only
  to the very shallow that syntax does not matter.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Perry E. Metzger
Subject: religion
Date: 
Message-ID: <87it1l2lup.fsf_-_@snark.piermont.com>
Erik Naggum <····@naggum.no> writes:
> * Perry E. Metzger
> | What I've noticed in a lot of people in this community (and it really
> | is just one community) is a lot of arrogance.
> 
>   Funny, that, I see a lot of humility and respect for the opinions and
>   thoughts of those older and wiser than oneself.  There is virtually no
>   arrogance among the cognoscenti.

When I asked one of the high cognoscs about the absence of an i/o
multiplexing primitive in his dialect because I needed one for doing
some event driven programming, I was regaled with why event driven
programming is stupid and ugly and I should be using threads
instead. I've had a dozen such experiences of late.

>   My experience has been one of great patience and willingness to
>   help those who wish to learn.  However, when some arrogant little
>   snot of a newbie walks into the forum only to spout his own
>   arrogant views without listening to anybody else, the /newbie/
>   will think this is arrogance.

I think calling anyone an "arrogant little snot" tends to lessen the
power of one's claim to humility, don't you?

> | I don't know about that.  I'd expect that if the community was
> | truly vibrant we'd be seeing things like equivalents to CPAN and such.
> 
>   I think this shows your arrogance more than anything else.  You
>   are obviously not willing to consider the fact that CPAN and the
>   like depend on much more than "vibrant communities" to exist.  In
>   particular, they depend on the comparative worthlessness of the
>   time spent on making software for their languages.

Ah, that explains it. Thank you for enlightening me.


-- 
Perry E. Metzger		·····@piermont.com
--
"Ask not what your country can force other people to do for you..."
From: Thomas Bushnell, BSG
Subject: Re: religion
Date: 
Message-ID: <87znuxi1bq.fsf@becket.becket.net>
"Perry E. Metzger" <·····@piermont.com> writes:

> When I asked one of the high cognoscs about the absence of an i/o
> multiplexing primitive in his dialect because I needed one for doing
> some event driven programming, I was regaled with why event driven
> programming is stupid and ugly and I should be using threads
> instead. I've had a dozen such experiences of late.

I think the cognoscs are right here.  Threads and multiplexing
primitives are duals of each other anyway, so if you are given
threads, you can easily create the multiplexing primitive you want.

But organizing design around non-blocking I/O primitives (from an OS
design standpoint) turns out to be much more complex and error prone
than having straightforward subroutine blocking primitives, and
telling users to use threads if they want multiplexing.

So that's why you get told that: it's actually the right thing.  If
the best way to represent your computation is with a non-blocking or a
multiplexing interface, then it's very easy to implement that using
the threads you've been given.
From: Perry E. Metzger
Subject: Re: religion
Date: 
Message-ID: <87bs7d114v.fsf@snark.piermont.com>
I promised someone that I was off this thread, but I will answer this
one last message on it.

·········@becket.net (Thomas Bushnell, BSG) writes:
> "Perry E. Metzger" <·····@piermont.com> writes:
> > When I asked one of the high cognoscs about the absence of an i/o
> > multiplexing primitive in his dialect because I needed one for doing
> > some event driven programming, I was regaled with why event driven
> > programming is stupid and ugly and I should be using threads
> > instead. I've had a dozen such experiences of late.
> 
> I think the cognoscs are right here.  Threads and multiplexing
> primitives are duals of each other anyway,

Actually, they aren't in practice -- there's a huge performance
difference. I could go into a long argument on why it is that event
driven programs are actually quite straightforward to write (they are
no more alien than functional programming is -- it is just a different
style), why they are much easier to write than thread driven programs
for certain applications (among other things, you end up with much
more elegant designs), and why in general the performance on real
hardware of event driven code is going to always be better than that
of thread driven code. I think this is all a digression though. We
aren't talking about threads vs. events. We're talking about an
attitude in the community.

The annoyance here is people assuming they know "the right thing" and
attempting to impose it, assuming they know more than the person
inquiring about his problem space. Maybe someone for whatever reason
needs to do something you don't particularly like -- work in an object
oriented style or a logic programming style or whatever style you
happen not to agree with this week. You aren't likely to be "right"
when you tell them "no that's stupid", you're just likely to be
expressing taste rather than "The Truth".

The Truth is that there isn't always a The Truth. Just because there
are problems with object systems doesn't mean objects are always
stupid. Just because you don't like someone writing a network protocol
that uses XML instead of sexprs doesn't mean that they don't have to
interface with something someone else built regardless of your
taste. Just because you might think there are no conditions in which
it might be better to do non-compacting garbage collection doesn't
mean someone might not have a condition in which he needs a
non-compacting collector, etc.

At best you can say "my experience is that it would be better for you
to use a hash table for this instead of a PATRICIA tree, but you can
implement the PATRICIA tree using structures this way..." and hell,
maybe it even turns out that the guy knows more about routing than you
do because he's a routing guy and you aren't and PATRICIA trees are
optimal for that problem.

> So that's why you get told that: it's actually the right thing.

I find in general that The Right Thing is hard to impose. I was
running a NetBSD related company for a couple of years. I found that
people want threads for some apps, and even though I didn't like
threads, I paid someone to work on a better performing threads
mechanism because some people find they're more convenient some of the
time than the techniques I prefer.

I could have told those people that they were idiots but they weren't
idiots -- they had reasons for wanting to use a technique other than
the one I found ideal and I couldn't really claim to understand their
problem domain as well as they did, so I shut up and worked on giving
them what they needed rather than what I felt The Right Thing was. I
had no idea if my notion of The Right Thing was right anyway. I am not
omniscient, and this is engineering, not mathematics.

In general, the languages and systems that win are the ones with the
least ideological axe to grind. I find that's true with people, too.


-- 
Perry E. Metzger		·····@piermont.com
--
"Ask not what your country can force other people to do for you..."
From: Ray Blaak
Subject: Re: religion
Date: 
Message-ID: <ubs7dx80w.fsf@telus.net>
"Perry E. Metzger" <·····@piermont.com> writes:
> The Truth is that there isn't always a The Truth.

Amen brother!

Seriously, hanging out on both c.l.l and c.l.s will quickly demonstrate that
the viritol level in the former is much much higher. Whether it is political
or religious is rather irrelevant, c.l.l is simply more intolerant.

--
Cheers,                                        The Rhythm is around me,
                                               The Rhythm has control.
Ray Blaak                                      The Rhythm is inside me,
·····@telus.net                                The Rhythm has my soul.
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240199350847640@naggum.no>
* Ray Blaak <·····@telus.net>
| Seriously, hanging out on both c.l.l and c.l.s will quickly demonstrate that
| the viritol level in the former is much much higher. Whether it is political
| or religious is rather irrelevant, c.l.l is simply more intolerant.

  Towards what?  We have more people in comp.lang.lisp who spend all their
  time complaining vociferously about how intolerant the newsgroup is, yet
  they have nothing whatsoever to communicate to anybody.  If you have any
  brilliant ideas for how to make sure that /nobody/ will respond to these
  cretins, please share your insight with you.  If, however, all you wish to
  do is complain, too, consider yourself part of the problem.  Sheesh.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Ray Blaak
Subject: Re: religion
Date: 
Message-ID: <u8z2gxvzu.fsf@telus.net>
Erik Naggum <····@naggum.no> writes:
> * Ray Blaak <·····@telus.net>
> | Seriously, hanging out on both c.l.l and c.l.s will quickly demonstrate that
> | the viritol level in the former is much much higher. Whether it is political
> | or religious is rather irrelevant, c.l.l is simply more intolerant.
> 
>   Towards what?  We have more people in comp.lang.lisp who spend all their
>   time complaining vociferously about how intolerant the newsgroup is, yet
>   they have nothing whatsoever to communicate to anybody.  If you have any
>   brilliant ideas for how to make sure that /nobody/ will respond to these
>   cretins, please share your insight with you.  If, however, all you wish to
>   do is complain, too, consider yourself part of the problem.  Sheesh.

Towards morons, differing points of view, Lisp 1, schemers, perl hackers,
copyleft, whatever.

I wasn't in fact complaining, only observing. Think about it: this thread is
*about* the tolerance levels of the newsgroups. The resulting discussion is
bound to have viewpoints of varying positions. It is disingenuous to simply
consider as invalid any position that is pointing out the problem.

The fix is not necessarily to make sure nobody responds to the cretins.  All
that is required is for people to lighten up, bigtime. That is, don't get
angry, upset, annoyed, etc., so easily. 

There are gentler ways with dealing with the clueless that are simply more
effective, namely, pointing out the error of their ways *without* invective
convulsions, and then if they don't get it, simply ignoring them or the entire
thread.

You yourself have improved dramatically in the last number of months, making
an obviously concerted effort to reign things in. I honestly congratulate you.

-- 
Cheers,                                        The Rhythm is around me,
                                               The Rhythm has control.
Ray Blaak                                      The Rhythm is inside me,
·····@telus.net                                The Rhythm has my soul.
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240231869826742@naggum.no>
* Ray Blaak
| I wasn't in fact complaining, only observing.

  Oh, come on!  You selectively "observe" and provide a complaint as a result
  because you rate your selective observations.  If you were truly /observing/,
  you would not feel such an enormous need to post a conclusion.

| Think about it: this thread is *about* the tolerance levels of the
| newsgroups.

  I have in fact observed that the vast majority of the idiotic fights in
  comp.lang.lisp are caused by some idiot who "observes" how hostile the
  newsgroup is -- to people who come to it to complain and do not even have
  the mental wherewithall to realize that they cannot both observe and comment
  negatively at the same time.  The two are mutually exclusive.  If you post
  your stupid observation, you elevate the things you have selected to some
  mystical importance where all the other observations that you probably did
  not even notice, vanish in the noise.  The fact is, /you/ highlight negative
  experiences and fit into the long line of people who have nothing to offer
  to this forum except your idiotic conclusions, /not/ observations, about the
  forum.  If we could find a way to get rid of these moronic meta-debates that
  people like you start and falsely accuse people of things you have not even
  bothered to pay attention to, your kind would not have much anything to
  complain about, either.

| It is disingenuous to simply consider as invalid any position that is
| pointing out the problem.

  But /why/ did you feel like posting your negative comment to begin with?

| The fix is not necessarily to make sure nobody responds to the cretins.  All
| that is required is for people to lighten up, bigtime. That is, don't get
| angry, upset, annoyed, etc., so easily.

  So do your part and don't piss people off with your selective "observations".

| There are gentler ways with dealing with the clueless that are simply more
| effective, namely, pointing out the error of their ways *without* invective
| convulsions, and then if they don't get it, simply ignoring them or the
| entire thread.

  Yeah, and I see your name trying to do this all the time, right?  You would
  not even  understand what you do until you have posted several thousand
  helpful messages to morons who want you to do their homework, to hold their
  hands, to answer their next question and the next because they are to fucking
  stupid to be able to figure out anything on their own.  Do this for a while
  and tell me that you follow your own advice.

| You yourself have improved dramatically in the last number of months, making
| an obviously concerted effort to reign things in.  I honestly congratulate you.

  Fuck you.  So why did you have to complain?  Goddamn whining negativists.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Ray Blaak
Subject: Re: religion
Date: 
Message-ID: <u3csoxpnm.fsf@telus.net>
Erik Naggum <····@naggum.no> writes:
> * Ray Blaak
> | The fix is not necessarily to make sure nobody responds to the cretins.  All
> | that is required is for people to lighten up, bigtime. That is, don't get
> | angry, upset, annoyed, etc., so easily.
> 
> So do your part and don't piss people off with your selective "observations".

My observations are hardly anything that would piss anyone off. Except you,
but everything does, so that's hardly anything to worry about.

> | There are gentler ways with dealing with the clueless that are simply more
> | effective, namely, pointing out the error of their ways *without* invective
> | convulsions, and then if they don't get it, simply ignoring them or the
> | entire thread.
> 
>   Yeah, and I see your name trying to do this all the time, right?

Google my name and you will quite readily see the style and tone of my
posts. Yes, I do in fact do this all the time.

> | You yourself have improved dramatically in the last number of months,
> | making an obviously concerted effort to reign things in.  I honestly
> | congratulate you.
> 
>   Fuck you.  So why did you have to complain?  Goddamn whining negativists.

For someone who is demonstrably so intelligent, your social skills suck big
time. A little of that intelligence applied to dealing with people would make
your communications dramatically more effective.

-- 
Cheers,                                        The Rhythm is around me,
                                               The Rhythm has control.
Ray Blaak                                      The Rhythm is inside me,
·····@telus.net                                The Rhythm has my soul.
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240240658198945@naggum.no>
* Ray Blaak
| My observations are hardly anything that would piss anyone off.  Except you,
| but everything does, so that's hardly anything to worry about.

  You just proved that your observational skills are clouded by emotion and
  hence are completely worthless.  Shut up and enjoy your leave of absence.

| For someone who is demonstrably so intelligent, your social skills suck big
| time. A little of that intelligence applied to dealing with people would make
| your communications dramatically more effective.

  Concern yourself with yourself.  If you have so much advice to offer me, how
  come you continue to insist on pissing me off more?  Obnoxius idiot.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Peter Keller
Subject: Re: religion
Date: 
Message-ID: <3d77886b$0$1128$80265adb@spool.cs.wisc.edu>
In comp.lang.scheme Erik Naggum <····@naggum.no> wrote:
:   Yeah, and I see your name trying to do this all the time, right?  You would
:   not even  understand what you do until you have posted several thousand
:   helpful messages to morons who want you to do their homework, to hold their
:   hands, to answer their next question and the next because they are to fucking
:   stupid to be able to figure out anything on their own.  Do this for a while
:   and tell me that you follow your own advice.

You, sir, need to walk away from computers, programming, and people
and get a different job in a different field where you don't even use
computers or talk to people because whatever computer/programming related
thing you are doing now is appearing to make you terribly unhappy.

-- 
-pete

E-mail address corrupted to stop spam.
Reply mail: psilord at cs dot wisc dot edu
I am responsible for what I say, noone else.
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240236311318050@naggum.no>
* Peter Keller
| You, sir, need to walk away from computers, programming, and people and get
| a different job in a different field where you don't even use computers or
| talk to people because whatever computer/programming related thing you are
| doing now is appearing to make you terribly unhappy.

  Let me know how it worked for you.  Take a week, nay, /four/ weeks off, and
  get back to us with a report on your mental state.  Until then, shut your hot
  air vent, you disgustingly rude little runt.  You speak about yourself and
  your inability to cope with rejection because you realize that you /should/
  be rejected for the insipid non-contributions you serve us.  Quit annoying
  people with your vacuous "observations" and your idiotic "advice".  OK?

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Peter Keller
Subject: Re: religion
Date: 
Message-ID: <3d77aa7e$0$1128$80265adb@spool.cs.wisc.edu>
In comp.lang.scheme Erik Naggum <····@naggum.no> wrote:
: * Peter Keller
: | You, sir, need to walk away from computers, programming, and people and get
: | a different job in a different field where you don't even use computers or
: | talk to people because whatever computer/programming related thing you are
: | doing now is appearing to make you terribly unhappy.

:   Let me know how it worked for you.  Take a week, nay, /four/ weeks off, and
:   get back to us with a report on your mental state.  Until then, shut your hot
:   air vent, you disgustingly rude little runt.  You speak about yourself and
:   your inability to cope with rejection because you realize that you /should/
:   be rejected for the insipid non-contributions you serve us.  Quit annoying
:   people with your vacuous "observations" and your idiotic "advice".  OK?

I think, from now on, I'm just going to read your posts instead of ever
responding to them.

-- 
-pete

E-mail address corrupted to stop spam.
Reply mail: psilord at cs dot wisc dot edu
I am responsible for what I say, noone else.
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240242105920727@naggum.no>
* Peter Keller
| I think, from now on, I'm just going to read your posts instead of ever
| responding to them.

  Naturally, this stupid trick would not have worked if you had just /done/ it.
  Take  a hike.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Paolo Amoroso
Subject: Re: religion
Date: 
Message-ID: <yI94PSZhi5ycshOT2H8vsXETJqKH@4ax.com>
[Original article crossposted to both comp.lang.scheme and comp.lang.lisp,
followup posted only to the latter]

On Thu, 05 Sep 2002 05:34:52 GMT, Ray Blaak <·····@telus.net> wrote:

> the viritol level in the former is much much higher. Whether it is political
> or religious is rather irrelevant, c.l.l is simply more intolerant.

For an example of what some technical forums demand from posters see:

  12 Steps to qmail List Bliss
  http://www.qcc.sk.ca/~charlesc/writings/12-steps-to-qmail-list-bliss.html

Incidentally, crossposting may not be the best way of reducing the amount
of vitriol.


Paolo
-- 
EncyCMUCLopedia * Extensive collection of CMU Common Lisp documentation
http://www.paoloamoroso.it/ency/README
From: Biep @ http://www.biep.org/
Subject: Re: religion
Date: 
Message-ID: <alau2o$1orceh$1@ID-63952.news.dfncis.de>
Ray Blaak wrote:
> Whether it is political or religious is rather
> irrelevant, c.l.l is simply more intolerant.

Well, Sonny, let me explain that to you.
You see, it is all about namespaces.  Common Lisp has two, and Scheme has
one.

Early in life every Schemer goes through exactly the same experience:
binding "list" to a list, and then calling it as a function.  Once you have
gone through that, you are "namespace aware", and are able to keep things
apart based on semantic domain.
(In fact the function "list" exists mainly for that purpose, Scheme being a
didactic language.)

Now, you see, Common Lispers are deprived of this experience, because their
function "list" will happily do what they want, despite the variable being
there.  That's because Common Lisp is foremost a production language.

Now you must know that Usenet, like Scheme but unlike Common Lisp, only has
a single namespace: while it may be a semantic error to discuss certain
things in certain corners, it is not a syntactic error.  In the situation
at hand, the two semantic domains are "people discussing programming
languages" and "programming languages".  Common Lispers tend to assume
these are stored in the "discussor cell" and the "discussee cell" of
variables, respectively

Now look what happens with a variable such as "you" in the following two
code samples:
* Correct usage is   'In article <····@CommonScheme> you wrote...'
* Incorrect usage is 'This shows that you are a braind-dead idiot'

Beware that on Usenet, both expressions will evaluate and not cause an
error.  The second one may cause an infinite loop, though, of the kind:
* No, you are one yourself
* That only proves you can't accept reality
* Look who's talking
  ...

Schemers are aware of this danger, and would not write code like the second
code sample above.  Common Lispers, on the other hand, are accustomed to
having the system make sure the "you" they are talking TO (= the one
addressed in the first code sample) is different from the "you" ahey are
talking ABOUT (= the one evaluated in the second code sample), and so end
up highly amazed that the side-effect of the code does' in fact, affect the
"you" in the discussor cell, and causes it to behave differently.

Moral: Whichever dialect is better, Scheme prepares one better for this
particular pitfall of Usenet.

[..and I suppose I'll need to include an explicit :-), because of another
rule of Usenet..]

--
Biep
Reply via any name whatsoever at the main domain corresponding to
http://www.biep.org
From: Ray Blaak
Subject: Re: religion
Date: 
Message-ID: <ubs7a5huw.fsf@telus.net>
"Biep @ http://www.biep.org/" <·········@my-web-site.com> writes:
> Ray Blaak wrote:
> > Whether it is political or religious is rather
> > irrelevant, c.l.l is simply more intolerant.
> 
> Well, Sonny, let me explain that to you.
> You see, it is all about namespaces.  Common Lisp has two, and Scheme has
> one.
[namespace/Scheme/CL/Usenet analogy elided]

Careful. These kinds of postings will piss people off.

A big honkin' :-) for the anger-management impaired.

--
Cheers,                                        The Rhythm is around me,
                                               The Rhythm has control.
Ray Blaak                                      The Rhythm is inside me,
·····@telus.net                                The Rhythm has my soul.
From: Daniel Barlow
Subject: Re: religion
Date: 
Message-ID: <87ofbck86l.fsf@noetbook.telent.net>
"Perry E. Metzger" <·····@piermont.com> writes:

> more elegant designs), and why in general the performance on real
> hardware of event driven code is going to always be better than that

For a minute there I thought we were into a potentially interesting
discussion of the different technical approaches to serving multiple
clients

> of thread driven code. I think this is all a digression though. We
> aren't talking about threads vs. events. We're talking about an
> attitude in the community.

But no, you have to drag the conversation back to _religion_.  Thank
you for your contribution to the c.l.l signal-to-noise ratio.


-dan

-- 

  http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources 
From: Kenny Tilton
Subject: Re: religion
Date: 
Message-ID: <3D781C79.1090100@nyc.rr.com>
Daniel Barlow wrote:
> "Perry E. Metzger" <·····@piermont.com> writes:
>>of thread driven code. I think this is all a digression though. We
>>aren't talking about threads vs. events. We're talking about an
>>attitude in the community.
> 
> 
> But no, you have to drag the conversation back to _religion_.  Thank
> you for your contribution to the c.l.l signal-to-noise ratio.
> 

Exactly. religion is the right subject, and what we have here is a 
missionary come to enlighten us. me, i love those stories where the 
heathens just burn the missionaries at the stake. in an NG, the only way 
to do that is to ignore them; these clowns feed on abuse.

kenny
From: Stephen J. Bevan
Subject: Re: religion
Date: 
Message-ID: <m3it1leyvl.fsf@dino.dnsalias.com>
·········@becket.net (Thomas Bushnell, BSG) writes:
> I think the cognoscs are right here.  Threads and multiplexing
> primitives are duals of each other anyway, so if you are given
> threads, you can easily create the multiplexing primitive you want.

In theory.  In practice I can't get a small Solaris, Linux, FreeBSD,
OpenBSD ... etc. box to run very well with 10,000 threads (one per
connection) but I can get it to run quite well with 10,000 connections
all being served by a single /dev/poll, kqueue or poll (ugh).  If
someone can make 10,000 threads run well on these platforms I'll give
up event driven approach in a flash.  I'm not holding my breath
though.
From: Thomas Bushnell, BSG
Subject: Re: religion
Date: 
Message-ID: <87lm68kbhr.fsf@becket.becket.net>
·······@dino.dnsalias.com (Stephen J. Bevan) writes:

> In theory.  In practice I can't get a small Solaris, Linux, FreeBSD,
> OpenBSD ... etc. box to run very well with 10,000 threads (one per
> connection) but I can get it to run quite well with 10,000 connections
> all being served by a single /dev/poll, kqueue or poll (ugh).  If
> someone can make 10,000 threads run well on these platforms I'll give
> up event driven approach in a flash.  I'm not holding my breath
> though.

That's a defect in those systems, and an excellent reason for those
systems having non-blocking I/O.  Not every system has such a defect,
however.  

Certainly any system which says "use threads, not non-blocking I/O"
should have cheap enough threads.
From: Duane Rettig
Subject: Re: religion
Date: 
Message-ID: <465xcd0l9.fsf@beta.franz.com>
·········@becket.net (Thomas Bushnell, BSG) writes:

> ·······@dino.dnsalias.com (Stephen J. Bevan) writes:
> 
> > In theory.  In practice I can't get a small Solaris, Linux, FreeBSD,
> > OpenBSD ... etc. box to run very well with 10,000 threads (one per
> > connection) but I can get it to run quite well with 10,000 connections
> > all being served by a single /dev/poll, kqueue or poll (ugh).  If
> > someone can make 10,000 threads run well on these platforms I'll give
> > up event driven approach in a flash.  I'm not holding my breath
> > though.
> 
> That's a defect in those systems, and an excellent reason for those
> systems having non-blocking I/O.  Not every system has such a defect,
> however.  

For 32-bit systems, I would characterize it as a memory limitation
and tradeoff, rather than a system defect.  If you consider what is
being asked for, 10,000 threads, which would each have a stack, could
only have a maximum individual stack size of less than 500 Kbytes
each (and that is assuming all 4 Gb is available, and that no other
memory is being used).  And 500 kbytes is pretty small - most Unix boxes
come with an average of 8 Mb stacks (except Tru64, which defaulted
to 2 Mb), and I am not sure about Windows defaults but they may be
in the same range.

> Certainly any system which says "use threads, not non-blocking I/O"
> should have cheap enough threads.

and enough resources.

-- 
Duane Rettig    ·····@franz.com    Franz Inc.  http://www.franz.com/
555 12th St., Suite 1450               http://www.555citycenter.com/
Oakland, Ca. 94607        Phone: (510) 452-2000; Fax: (510) 452-0182   
From: Tim Bradshaw
Subject: Re: religion
Date: 
Message-ID: <ey3k7lsec3m.fsf@cley.com>
* Duane Rettig wrote:

> For 32-bit systems, I would characterize it as a memory limitation
> and tradeoff, rather than a system defect.  If you consider what is
> being asked for, 10,000 threads, which would each have a stack, could
> only have a maximum individual stack size of less than 500 Kbytes
> each (and that is assuming all 4 Gb is available, and that no other
> memory is being used).  And 500 kbytes is pretty small - most Unix boxes
> come with an average of 8 Mb stacks (except Tru64, which defaulted
> to 2 Mb), and I am not sure about Windows defaults but they may be
> in the same range.

On Solaris/SPARC at least, it looks as if the default pthread stack
size is 1 or 2MB, but you can set it in bytes and so could set it to
be quite small.  However I haven't actually *tried* this - it may not
actually save any space or not work at all.  And I don't know what
Java does, either.

I think that your point is still a good one - 10,000 threads is till
really a lot of threads in terms of resources unless threads can be
made incredibly lightweight.

It would also be interesting to know how badly systems behave with
this many threads - I suspect lots of systems (but probably not
big-iron Unixes) have schedulers which don't behave too well with that
many many things to schedule.

--tim
From: Will Deakin
Subject: Re: religion
Date: 
Message-ID: <alo8de$t3g$1@knossos.btinternet.com>
Tim wrote:
> On Solaris/SPARC at least, it looks as if the default pthread stack
> size is 1 or 2MB, but you can set it in bytes and so could set it to
> be quite small.
Hmmm. I'm not sure about this although I have probably completely 
misunderstood what is going on.

I belive that on Solaris/SPARC 7+ most of the process state is shared 
with the LWP -- that is address space, stack and so on. With 2.6 the 
stack had a soft limit of 8MB and a hard limit of 2GB. On later releases 
the hard limit is `infinite.' Each user/pthread has a pointer into the 
stack so that for, say 10 000 pthreads, by default the pthread will have 
about between 8KB and no upper bound, assuming you have enough memory.

> It would also be interesting to know how badly systems behave with
> this many threads - I suspect lots of systems (but probably not
> big-iron Unixes) have schedulers which don't behave too well with that
> many many things to schedule.
The number of LWP's bound to a process -- and thus affecting the chances 
of a pthread running -- can be upped. However, even with a meaty 128 CPU 
E15k you are still looking at 80 pthreads a cpu -- potentially all 
sheduled via on process. Ouch.

:)w
From: Will Deakin
Subject: Re: religion
Date: 
Message-ID: <alo8je$t3g$2@knossos.btinternet.com>
I wrote:
> sheduled via on process. Ouch.
Make that: "sheduled via one process. Ouch."

:(
From: Hannah Schroeter
Subject: Re: religion
Date: 
Message-ID: <aloc9c$blh$1@c3po.schlund.de>
Hello!

Will Deakin  <···········@hotmail.com> wrote:

>I belive that on Solaris/SPARC 7+ most of the process state is shared 
>with the LWP -- that is address space, stack and so on.

How do you share the stack if different LWPs are to have different
threads of execution, i.e. different function call stacks?

>[...]

Kind regards,

Hannah.
From: Duane Rettig
Subject: Re: religion
Date: 
Message-ID: <4y9a8xgnk.fsf@beta.franz.com>
······@schlund.de (Hannah Schroeter) writes:

> Hello!
> 
> Will Deakin  <···········@hotmail.com> wrote:
> 
> >I belive that on Solaris/SPARC 7+ most of the process state is shared 
> >with the LWP -- that is address space, stack and so on.
> 
> How do you share the stack if different LWPs are to have different
> threads of execution, i.e. different function call stacks?

It may be possible to map different physical memory into the same
virtual address ranges, thereby getting around my concern.  I am
not sure if anyone does this, though, and I am not sure how
lightweight such threads would actually be.

-- 
Duane Rettig    ·····@franz.com    Franz Inc.  http://www.franz.com/
555 12th St., Suite 1450               http://www.555citycenter.com/
Oakland, Ca. 94607        Phone: (510) 452-2000; Fax: (510) 452-0182   
From: Will Deakin
Subject: Re: religion
Date: 
Message-ID: <alpj5v$lvi$1@newsreaderm1.core.theplanet.net>
Duane Rettig wrote:
> It may be possible to map different physical memory into the same
> virtual address ranges, thereby getting around my concern.
AFAIUI that is what happens.

> I am not sure if anyone does this, though, and I am not sure how
> lightweight such threads would actually be.
Not very. The whole thing about LWP's are that they are the native 
threads running processes in the OS. There is a thin smear of 
interface and, plop, like a rabbit from a hat you have a fully fledged 
process. Not very lightweight at all, really.

:)w
From: Hannah Schroeter
Subject: Thread implementation overheads (was Re: religion)
Date: 
Message-ID: <am7b07$uhm$1@c3po.schlund.de>
Hello!

Duane Rettig  <·····@franz.com> wrote:

>> >I belive that on Solaris/SPARC 7+ most of the process state is shared 
>> >with the LWP -- that is address space, stack and so on.

>> How do you share the stack if different LWPs are to have different
>> threads of execution, i.e. different function call stacks?

>It may be possible to map different physical memory into the same
>virtual address ranges, thereby getting around my concern.

Of course, you can. Usually, you share the code and non-stack
data segment(s) in threads, while have different memory for the
stack. In fact, when creating a thread, you can also lazily copy
it (i.e. copy-on-write). However, with stack based activation frames,
you'll pay at least one page of memory per thread, e.g. 4kb on x86.
With a non-stack-based execution model, things may be different.
Erlang, for example, achieves cheaper threads (called processes in
Erlang), with a memory "overhead" (base cost) of less than 1 kb per
thread. GHC 4's threads cost about 1 to 2 kb per thread as base cost.
I haven't really measured SML/NJ's thread overhead, but I expect it
to be quite low, too, as SML/NJ does automatic CPS transform and
uses continuation-switching as implementation strategy for the thread
abstractions.

>I am
>not sure if anyone does this, though, and I am not sure how
>lightweight such threads would actually be.

See above. For all of these, I haven't measured thread descriptors or
such.

Kind regards,

Hannah.
From: Hans Boehm
Subject: Re: Thread implementation overheads (was Re: religion)
Date: 
Message-ID: <am7soa$7vv$1@hplms2.hpl.hp.com>
"Hannah Schroeter" <······@schlund.de> wrote in message
·················@c3po.schlund.de...
> Hello!
>
> Duane Rettig  <·····@franz.com> wrote:
>
> >> >I belive that on Solaris/SPARC 7+ most of the process state is shared
> >> >with the LWP -- that is address space, stack and so on.
>
> >> How do you share the stack if different LWPs are to have different
> >> threads of execution, i.e. different function call stacks?
>
> >It may be possible to map different physical memory into the same
> >virtual address ranges, thereby getting around my concern.
>
> Of course, you can. Usually, you share the code and non-stack
> data segment(s) in threads, while have different memory for the
> stack. In fact, when creating a thread, you can also lazily copy
> it (i.e. copy-on-write). However, with stack based activation frames,
> you'll pay at least one page of memory per thread, e.g. 4kb on x86.
> With a non-stack-based execution model, things may be different.
> Erlang, for example, achieves cheaper threads (called processes in
> Erlang), with a memory "overhead" (base cost) of less than 1 kb per
> thread. GHC 4's threads cost about 1 to 2 kb per thread as base cost.
> I haven't really measured SML/NJ's thread overhead, but I expect it
> to be quite low, too, as SML/NJ does automatic CPS transform and
> uses continuation-switching as implementation strategy for the thread
> abstractions.
>
As a practical matter, you may not want to do this, for two distinct
reasons:

1. I'm told that there is a significant advantage in the kernel if you are
able to share the ENTIRE address space between threads.  You can use the
same data structure to represent the address mapping for all the threads,
and presumably you don't need to worry about invalidating TLB entries etc.
if you switch between them.  Some operating systems already lose this
advantage for other reasons, but others (AFAIK, Linux is among them) do
share the entire address space.

2. For C/C++ programs, it makes sense for one thread to refer to another
thread's stack.  (E.g. one threads builds up some data structure describing
a problem on its stack, creates N threads to help solve the problem, giving
them each a pointer to the stack-allocated data structure, waits for them to
finish, and returns, implicitly popping the problem description from the
stack.)  All C/C++ thread implementations of which I'm aware allow this.
This doesn't directly matter for Scheme (or Java) unless you have an
exceptionally clever compiler to introduce this sort of thing.  On the other
hand, there are large advantages to using the system-standard threading
library.  For example, it otherwise becomes very difficult to interoperate
with third party libraries that also need to create threads.

As Rob Warnock pointed out, there is a real tradeoff here.

Hans
From: Hannah Schroeter
Subject: Re: Thread implementation overheads (was Re: religion)
Date: 
Message-ID: <ao0ogh$c27$1@c3po.schlund.de>
Hello!

Hans Boehm <··········@hp.com> wrote:

>[...]

>> Of course, you can. Usually, you share the code and non-stack
>> data segment(s) in threads, while have different memory for the
>> stack. In fact, when creating a thread, you can also lazily copy
>> it (i.e. copy-on-write). However, with stack based activation frames,
>> you'll pay at least one page of memory per thread, e.g. 4kb on x86.
>> With a non-stack-based execution model, things may be different.
>> Erlang, for example, achieves cheaper threads (called processes in
>> Erlang), with a memory "overhead" (base cost) of less than 1 kb per
>> thread. GHC 4's threads cost about 1 to 2 kb per thread as base cost.
>> I haven't really measured SML/NJ's thread overhead, but I expect it
>> to be quite low, too, as SML/NJ does automatic CPS transform and
>> uses continuation-switching as implementation strategy for the thread
>> abstractions.

>As a practical matter, you may not want to do this, for two distinct
                                               ^^^^
>reasons:

That "this" is not really defined. What do you mean?
- Allocating stack on the same virtual addresses, making the stack
  of the threads unaccessible to others (except for the lazy copy
  created on thread creation time)
- CPS transform?

>1. I'm told that there is a significant advantage in the kernel if you are
>able to share the ENTIRE address space between threads.  You can use the
>same data structure to represent the address mapping for all the threads,
>and presumably you don't need to worry about invalidating TLB entries etc.
>if you switch between them.  Some operating systems already lose this
>advantage for other reasons, but others (AFAIK, Linux is among them) do
>share the entire address space.

It depends on whether you request that sharing, of course, and whether
the OS has the special case detection ("oh, memory map exactly the
same, I can omit the reload of the address translation control register,
the TLB, and the cache flushes [the latter for caches which are indexed
by logical addresses]").

However, complete sharing of the memory map does NOT reduce the memory
overhead of *at least* one pre-allocated page for each thread's stack,
on x86 being 4KB. Perhaps another page or two for the kernel level
process specific stack, as that probably can't be shared either.

>2. For C/C++ programs, it makes sense for one thread to refer to another
>thread's stack.  (E.g. one threads builds up some data structure describing
>a problem on its stack, creates N threads to help solve the problem, giving
>them each a pointer to the stack-allocated data structure, waits for them to
>finish, and returns, implicitly popping the problem description from the
>stack.)  All C/C++ thread implementations of which I'm aware allow this.

As long as the structure is read-only, this still would work for a
copy-on-write stack. But I see this point.

>This doesn't directly matter for Scheme (or Java) unless you have an
>exceptionally clever compiler to introduce this sort of thing.  On the other
>hand, there are large advantages to using the system-standard threading
>library.  For example, it otherwise becomes very difficult to interoperate
>with third party libraries that also need to create threads.

>As Rob Warnock pointed out, there is a real tradeoff here.

>Hans

There are many tradeoffs in fact. And they do depend on the particular
application, too. Do you use threads to simplify concurrent network
I/O (a userland scheduling thread library, perhaps even without timeslice
based preemption, instead only preempting on blocking network I/O calls,
could suffice), do you use them to distribute CPU bound work (an 1:1
or M:N thread library that uses at least #CPUs kernel level
threads/processes probably would be better), etc.?

Kind regards,

Hannah.
From: Hans-J. Boehm
Subject: Re: Thread implementation overheads (was Re: religion)
Date: 
Message-ID: <1178a29f.0210161505.71fccc13@posting.google.com>
[My apologies if this is a repeat.  My first posting attempt doesn't
appear to have made it everywhere.]

"Hannah Schroeter" <······@schlund.de> wrote in message
·················@c3po.schlund.de...
> Hello!
>
> ...
> That "this" is not really defined. What do you mean?
> - Allocating stack on the same virtual addresses, making the stack
>   of the threads unaccessible to others (except for the lazy copy
>   created on thread creation time)
> - CPS transform?
I was really talking about the former, which forces address spaces in
separate threads to differ.  I'm sorry I didn't make that clear.

However, as I hinted at in the second point below, I fundamentally
believe
that practical systems generally need to standardize on a single
threads
implementation, which would limit the role of explicit
continutation-switching implementations.  We're rapidly getting to the
point
where the average new desktop machine has more than one logical
processor.
In order to take advantage of this, standard low-level libraries will
probably need to be able to take advantage of threads. And those
threads
have to cooperate with the threads introduced by the higher layers. 
That's
very hard unless they're the same.  (Some Java implementors have
argued that
it's not impossible.  I think both Jikes and NaturalBridge Java
implementations try quite hard to have their own threads interact
correctly
with underlying system-provided threads.  I'm not sure the payoff will
continue to justify the complexity as system threads implementations
improve.)

> >1. I'm told that there is a significant advantage in the kernel if you
are
> >able to share the ENTIRE address space between threads.
...
> It depends on whether you request that sharing, of course, and whether
> the OS has the special case detection ("oh, memory map exactly the
> same, I can omit the reload of the address translation control register,
> the TLB, and the cache flushes [the latter for caches which are indexed
> by logical addresses]").
>
> However, complete sharing of the memory map does NOT reduce the memory
> overhead of *at least* one pre-allocated page for each thread's stack,
> on x86 being 4KB. Perhaps another page or two for the kernel level
> process specific stack, as that probably can't be shared either.
Agreed.  Complete sharing increases the virtual address space
requirements,
and probably doesn't affect the physical memory requirements.  Yet
another
reason to move to 64-bit addressing ...
>
> >2. For C/C++ programs, it makes sense for one thread to refer to another
> >thread's stack.  (E.g. one threads builds up some data structure
describing
> >a problem on its stack, creates N threads to help solve the problem,
giving
> >them each a pointer to the stack-allocated data structure, waits for them
to
> >finish, and returns, implicitly popping the problem description from the
> >stack.)  All C/C++ thread implementations of which I'm aware allow this.
>
> As long as the structure is read-only, this still would work for a
> copy-on-write stack. But I see this point.
Given that this is C, I would expect a typical use to include a result
buffer on the original thread's stack.  So it  wouldn't always be
read-only.
But that's still an interesting observation.
...
>
> There are many tradeoffs in fact. And they do depend on the particular
> application, too. Do you use threads to simplify concurrent network
> I/O (a userland scheduling thread library, perhaps even without timeslice
> based preemption, instead only preempting on blocking network I/O calls,
> could suffice), do you use them to distribute CPU bound work (an 1:1
> or M:N thread library that uses at least #CPUs kernel level
> threads/processes probably would be better), etc.?
>
Assuming that we are not talking about an embedded application, my
guess is
that sooner or later you will probably want to do all of the above,
plus
using threads to simplify GUI programming, etc.  Possibly it will all
be in
the same application.  Even if it isn't, you'll want to standardize,
so that
libraries only need to support one kind of threads and
synchronization.
That means you really want a 1:1 or M:N fully preemptive and
time-sliced
thread implementation.  (I'd argue fairly strongly for 1:1, but I
think I
already made enough controversial statements for one message.)

Hans
From: Rob Warnock
Subject: Re: religion
Date: 
Message-ID: <uo06q2m2mcit08@corp.supernews.com>
Hannah Schroeter <······@schlund.de> wrote:
+---------------
| Will Deakin  <···········@hotmail.com> wrote:
| >I belive that on Solaris/SPARC 7+ most of the process state is shared 
| >with the LWP -- that is address space, stack and so on.
| 
| How do you share the stack if different LWPs are to have different
| threads of execution, i.e. different function call stacks?
+---------------

Seems like one ought to be able to use copy-on-write on the stack segment,
wherein the sharing breaks when a thread extends the stack, while doing
ordinary sharing on the text and data segments (including shared libs).

This *would* have the limitation that one thread would not be able to
access the stack of another thread, so that all interthread communication
would have to be done through the data segment. Is that acceptable?
Your call...


-Rob

-----
Rob Warnock, PP-ASEL-IA		<····@rpw3.org>
627 26th Avenue			<URL:http://www.rpw3.org/>
San Mateo, CA 94403		(650)572-2607
From: Will Deakin
Subject: Re: religion
Date: 
Message-ID: <alpgk2$nig$1@venus.btinternet.com>
Hannah Schroeter wrote:
>>I belive that on Solaris/SPARC 7+ most of the process state is shared 
>>with the LWP -- that is address space, stack and so on.
> 
> How do you share the stack if different LWPs are to have different
> threads of execution, i.e. different function call stacks?
I think there is a misunderstanding here. The LWP -- in solaris at least 
-- is the entity that is bound to the kernel thread (a.k.a. kthread). It 
is this binding process that causes the LWP to be scheduled for cpu time 
and as such the LWP is the key consitutent of a process.

There is then a whole stack (sic) of stuff that need to be considered. 
Like increasing the number of LWP's that a process has and how the 
kernel scheduler interacts with the process and user- a.k.a. pthread 
scheduler. And how multiple LWP's share memory.

:)w
From: Stephen J. Bevan
Subject: Re: religion
Date: 
Message-ID: <m3u1kwzf0r.fsf@dino.dnsalias.com>
·········@becket.net (Thomas Bushnell, BSG) writes:
> That's a defect in those systems, and an excellent reason for those
> systems having non-blocking I/O.  Not every system has such a defect,
> however.  

Maybe not, but those are the systems I have to work with.

> Certainly any system which says "use threads, not non-blocking I/O"
> should have cheap enough threads.

Perhaps but none of those systems say that.  If you happen to know of
one that has a complete IPsec stack and can run Common Lisp or Scheme
then I'm interested in hearing about it.
From: Luke Gorrie
Subject: Re: religion
Date: 
Message-ID: <lhd6rk5sct.fsf@bluetail.com>
·········@becket.net (Thomas Bushnell, BSG) writes:

> ·······@dino.dnsalias.com (Stephen J. Bevan) writes:
> 
> > In theory.  In practice I can't get a small Solaris, Linux, FreeBSD,
> > OpenBSD ... etc. box to run very well with 10,000 threads (one per
> > connection) but I can get it to run quite well with 10,000 connections
> > all being served by a single /dev/poll, kqueue or poll (ugh).  If
> > someone can make 10,000 threads run well on these platforms I'll give
> > up event driven approach in a flash.  I'm not holding my breath
> > though.

[Replying to a reply to the actual message, can't find the original]

If you haven't looked at Erlang, you might get a real kick out of
it.

Erlang has light-weight processes, and programming them is similar to
threads or unix processes. In particular, they can use blocking I/O. I
think a process uses about 1KB of memory initially, and their stacks
are dynamically sized.

The runtime system is a C program with its own scheduler for Erlang
processes. It uses non-blocking I/O in an event loop based on some
poll()-like function such as (depending on OS) poll, kqueue, or kpoll
[1].

So you basically get the scalability of an event-driven system, and
the convenience of high-level process abstractions with blocking
operations. It's very well tried and proven in commercial networking
products, and really a joy to use.

[1]: The guys who do Erlang's networking have implemented a fast
'/dev/kpoll' mechanism as a Linux kernel extension:
http://www.synap.se/open_source.html. In his announcement to the
erlang-questions mailing list, Per Bergqvist reported running 1MB/sec
across 8192 TCP sockets in an Erlang program running on a Celeron 650,
with 70% idle CPU remaining. IIRC there was a separate process for
each socket and the machines were echoing data to each other.

NB: The latest release of Erlang has support for kpoll as a configure
option, so you only need to patch the kernel, not the Erlang system.

Cheers,
Luke
From: Stephen J. Bevan
Subject: Re: religion
Date: 
Message-ID: <m3bs74z5vm.fsf@dino.dnsalias.com>
Luke Gorrie <····@bluetail.com> writes:
> If you haven't looked at Erlang, you might get a real kick out of
> it.

I have.  I think it is very interesting, however I didn't use it because ...

> The runtime system is a C program with its own scheduler for Erlang
> processes. It uses non-blocking I/O in an event loop based on some
> poll()-like function such as (depending on OS) poll, kqueue, or kpoll
> [1].

I built something similar in Scheme using call/cc (+ kqueue, poll, or
/dev/poll) .  I've also done it in Common Lisp and Scheme without
call/cc by not pretending to have threads at all and just writing in
CPS.  That's not to say I would not consider using Erlang in the
future since it includes lots of other features (I'm particularly
interested in the database).  However, as with CL/Scheme, selling any
language other than C/C++/Java to other developers is tough (they
worry that this "unknown" language won't help them get another job),
let alone to management (who tend to worry where they will find
programmers for this "unknown" language).

> NB: The latest release of Erlang has support for kpoll as a configure
> option, so you only need to patch the kernel, not the Erlang system.

"only need to patch the kernel" is not something that goes down well
with some customers.  I have experience of that when telling them that
they "only need to patch the kernel" in order to install FreeS/WAN
(this is before they made a module version available) on their RedHat
boxes.  Hopefully Linus or a vendor will include /dev/kpoll or
something like it soon so that patching is not required.
From: Luke Gorrie
Subject: Re: religion
Date: 
Message-ID: <lhy9a83yop.fsf@bluetail.com>
·······@dino.dnsalias.com (Stephen J. Bevan) writes:

> > NB: The latest release of Erlang has support for kpoll as a configure
> > option, so you only need to patch the kernel, not the Erlang system.
> 
> "only need to patch the kernel" is not something that goes down well
> with some customers.  I have experience of that when telling them that
> they "only need to patch the kernel" in order to install FreeS/WAN
> (this is before they made a module version available) on their RedHat
> boxes.  Hopefully Linus or a vendor will include /dev/kpoll or
> something like it soon so that patching is not required.

This is perhaps the best reason for selling boxes instead of
programs. :-)

Cheers,
Luke
From: Stephen J. Bevan
Subject: Re: religion
Date: 
Message-ID: <m3k7lsxeyg.fsf@dino.dnsalias.com>
Luke Gorrie <····@bluetail.com> writes:
> ·······@dino.dnsalias.com (Stephen J. Bevan) writes:
> > "only need to patch the kernel" is not something that goes down well
> > with some customers.  I have experience of that when telling them that
> > they "only need to patch the kernel" in order to install FreeS/WAN
> > (this is before they made a module version available) on their RedHat
> > boxes.  Hopefully Linus or a vendor will include /dev/kpoll or
> > something like it soon so that patching is not required.
> 
> This is perhaps the best reason for selling boxes instead of
> programs. :-)

The most memorable thing I can remember about one CEO I worked for was
his statement that "in 23 years in the computer business, one thing
I've learned is it is tough to make money selling hardware to run the
software."  That came in response to a question about the future of a
VPN product that included hardware :-)
From: Aleksandr Skobelev
Subject: Re: religion
Date: 
Message-ID: <m3sn0foh2v.fsf@list.ru>
·······@dino.dnsalias.com (Stephen J. Bevan) writes:

> ·········@becket.net (Thomas Bushnell, BSG) writes:
> > I think the cognoscs are right here.  Threads and multiplexing
> > primitives are duals of each other anyway, so if you are given
> > threads, you can easily create the multiplexing primitive you want.
> 
> In theory.  In practice I can't get a small Solaris, Linux, FreeBSD,
> OpenBSD ... etc. box to run very well with 10,000 threads (one per
> connection) but I can get it to run quite well with 10,000 connections
> all being served by a single /dev/poll, kqueue or poll (ugh).  If
> someone can make 10,000 threads run well on these platforms I'll give
> up event driven approach in a flash.  I'm not holding my breath
> though.

Could you give some details, please? 
How small were those boxes? :) What system parameters in Linux and
FreeBSD did you have to set to poll on 10,000 sockets? Why did you use
poll but not select? Is it reasonable to use the same approach on
Windows (with select())? 

TIA 
From: Stephen J. Bevan
Subject: Re: religion
Date: 
Message-ID: <m38z27xjgc.fsf@dino.dnsalias.com>
Aleksandr Skobelev <···········@list.ru> writes:
> Could you give some details, please? 

I could but I'd be rehashing a lot of what has already been said in
much more detail at http://www.kegel.com/c10k.html
From: Aleksandr Skobelev
Subject: Re: religion
Date: 
Message-ID: <m3sn0f87ei.fsf@list.ru>
·······@dino.dnsalias.com (Stephen J. Bevan) writes:

> Aleksandr Skobelev <···········@list.ru> writes:
> > Could you give some details, please? 
> 
> I could but I'd be rehashing a lot of what has already been said in
> much more detail at http://www.kegel.com/c10k.html

Thank you. I'll read it. It looks very useful.
From: Sander Vesik
Subject: Re: religion
Date: 
Message-ID: <1031937528.681638@haldjas.folklore.ee>
In comp.lang.scheme Aleksandr Skobelev <···········@list.ru> wrote:
> ·······@dino.dnsalias.com (Stephen J. Bevan) writes:
> 
>> ·········@becket.net (Thomas Bushnell, BSG) writes:
>> > I think the cognoscs are right here.  Threads and multiplexing
>> > primitives are duals of each other anyway, so if you are given
>> > threads, you can easily create the multiplexing primitive you want.
>> 
>> In theory.  In practice I can't get a small Solaris, Linux, FreeBSD,
>> OpenBSD ... etc. box to run very well with 10,000 threads (one per
>> connection) but I can get it to run quite well with 10,000 connections
>> all being served by a single /dev/poll, kqueue or poll (ugh).  If
>> someone can make 10,000 threads run well on these platforms I'll give
>> up event driven approach in a flash.  I'm not holding my breath
>> though.
> 
> Could you give some details, please? 
> How small were those boxes? :) What system parameters in Linux and
> FreeBSD did you have to set to poll on 10,000 sockets? Why did you use
> poll but not select? Is it reasonable to use the same approach on
> Windows (with select())? 

select is implemeted as a thing wrapper around poll these days on most
modern unix systems. 

> 
> TIA 

-- 
	Sander

+++ Out of cheese error +++
From: Peter Keller
Subject: Re: religion
Date: 
Message-ID: <3d76d560$0$1125$80265adb@spool.cs.wisc.edu>
In comp.lang.scheme Perry E. Metzger <·····@piermont.com> wrote:

: Erik Naggum <····@naggum.no> writes:
:> * Perry E. Metzger
:> | I don't know about that.  I'd expect that if the community was
:> | truly vibrant we'd be seeing things like equivalents to CPAN and such.
:> 
:>   I think this shows your arrogance more than anything else.  You
:>   are obviously not willing to consider the fact that CPAN and the
:>   like depend on much more than "vibrant communities" to exist.  In
:>   particular, they depend on the comparative worthlessness of the
:>   time spent on making software for their languages.

It is statements like these that cause me to pick another popular C
library to write an FFI interface for in the scheme implementation
I've chosen.

I'm beginning to believe that the reason scheme isn't as popular as the
rest of the languages isn't so much as there are many implementations, but
more so that no-one understands how to market it worth a damn.

Seen any oreilly books on scheme, or scheme interfaces to SQL or anything like
that? My point exactly.

-- 
-pete

E-mail address corrupted to stop spam.
Reply mail: psilord at cs dot wisc dot edu
I am responsible for what I say, noone else.
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240187696442627@naggum.no>
* Peter Keller <·······@data.upl.cs.wisc.edu>
| It is statements like these that cause me to pick another popular C library
| to write an FFI interface for in the scheme implementation I've chosen.

  No, it is not.  If you had understood the word "comparative", you would not
  have been able to use my statements to support your preconceived notions.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Peter Keller
Subject: Re: religion
Date: 
Message-ID: <3d76e307$0$1125$80265adb@spool.cs.wisc.edu>
In comp.lang.scheme Erik Naggum <····@naggum.no> wrote:
: * Peter Keller <·······@data.upl.cs.wisc.edu>
: | It is statements like these that cause me to pick another popular C library
: | to write an FFI interface for in the scheme implementation I've chosen.

:   No, it is not.  If you had understood the word "comparative", you would not
:   have been able to use my statements to support your preconceived notions.

SunOS buzzard > webster comparative
 1 com.par.a.tive adj  \k*m-'par-*t-iv\
      1 : of, relating to, or constituting the degree of comparison in 
          a language that denotes increase in the quality, quantity, or 
          relation expressed by an adjective or adverb
      2 : considered as if in comparison to something else as a 
          standard not quite attained : RELATIVE [~ 
          stranger]
      3 : studied systematically by comparison of phenomena [~ 
          literature]
    com.par.a.tive.ly adv 
    com.par.a.tive.ness n 
 2 comparative n 
      1 : one that compares with another esp. on equal footing 
          : RIVAL; specif : one that makes witty 
          or mocking comparisons
      2 : the comparative degree or form in a language

I'm sorry, but I guess I just don't understand how you wanted to use it in
the context with "worthless". 

-- 
-pete

E-mail address corrupted to stop spam.
Reply mail: psilord at cs dot wisc dot edu
I am responsible for what I say, noone else.
From: Christopher Browne
Subject: Re: religion
Date: 
Message-ID: <al6mtn$1njij4$2@ID-125932.news.dfncis.de>
In the last exciting episode, Peter Keller <·······@data.upl.cs.wisc.edu> wrote::
> I'm beginning to believe that the reason scheme isn't as popular as
> the rest of the languages isn't so much as there are many
> implementations, but more so that no-one understands how to market
> it worth a damn.

> Seen any oreilly books on scheme, or scheme interfaces to SQL or
> anything like that? My point exactly.

O'Reilly also doesn't have books on Common Lisp, and aren't accepting
new titles on TeX or _any_ form of Lisp, so I'd not count that as
being of much interest.

And the following bit of Guile code does _exactly_ what it would be
expected to do, namely interface with PostgreSQL, and pull a set of
entries.

(use-modules (database postgres))
(define test (pg-connectdb "dbname=sqlledger"))
(define result (pg-exec test "select * from zipcode"))
(let loop ((index 0))
   (if (< index (pg-ntuples result))
       (begin
	  (display (pg-getvalue result index 0))
	  (display (pg-getvalue result index 1))
	  (display (pg-getvalue result index 2))
	  (display (pg-getvalue result index 3))
	  (display (pg-getvalue result index 4))
	  (newline)
	  (loop (+ index 1)))))

The PG binding is a little more "primitive" than you'd usually use in
CL, basically eschewing all use of macros, and not really providing
any help with transactions/cursors (both being places where some
well-place macros would be _really_ useful).

But it exists, and it works, and I just used it to blow up an Emacs
buffer by showing 80,000-odd zipcodes.
-- 
(concatenate 'string "aa454" ·@freenet.carleton.ca")
http://cbbrowne.com/info/oses.html
Coming  Soon  to a  Mainframe  Near  You!   MICROS~1 Windows  NT  6.0,
complete with VISUAL JCL...
From: Peter Keller
Subject: Re: religion
Date: 
Message-ID: <3d76f274$0$1128$80265adb@spool.cs.wisc.edu>
In comp.lang.scheme Christopher Browne <········@acm.org> wrote:
: In the last exciting episode, Peter Keller <·······@data.upl.cs.wisc.edu> wrote::
:> I'm beginning to believe that the reason scheme isn't as popular as
:> the rest of the languages isn't so much as there are many
:> implementations, but more so that no-one understands how to market
:> it worth a damn.

:> Seen any oreilly books on scheme, or scheme interfaces to SQL or
:> anything like that? My point exactly.

: O'Reilly also doesn't have books on Common Lisp, and aren't accepting
: new titles on TeX or _any_ form of Lisp, so I'd not count that as
: being of much interest.

Well, I can buy an oreilly book on "mastering regular expressions" but not
"mastering macros". Why? Why doesn't oreilly accept books for lisp-like
lanauges, and why isn't there a book out there like a "macro cookbook"
which talks about the various forms of macro representation and what
you can do with them with good *and relevant* examples by any publisher
at all? The regex book pulled it off and there are a BUNCH of regular
expression styles out there that are incompatible.

: And the following bit of Guile code does _exactly_ what it would be
: expected to do, namely interface with PostgreSQL, and pull a set of
: entries.

I wasn't talking about implementations. I was talking about why I can
wander to a book store and see shelf upon shelf of ASP/JAVA/PERL/C++/SQL
books and then find exactly one grungy SICP book mislocated into the
COBOL section--the only scheme book to be found.

<rant>
You want religion? I'll give you religion:

So, I'm sitting at work and my office mate says he wants to learn scheme
to implement a project that scheme would be a perfect match for. So,
I get to helping him learn it, but he is obviously shy about it. So,
I ask him about it and do you know what he says?

"Scheme and Lisp, in general, are pretentious. It tries to be a magic
bullet which attempts to solve all problems but the solutions it comes
up with (everything is a heterogeneous list, strange typechecking, wierd
development environment, strange/bad error reporting) just get in the
way of doing "real work". Also, they think their way is the "right and
only way" and that just bothers me."

I didn't know how to answer him other than some fast excuse that scheme is
just like any other language, you just have to think a little differently
to use its full potential. To which he scowled at me, as if I had made
true his thoughts about the subject, but continued to learn it because
scheme really WAS the best solution for his problem.  
</rant>

It is this kind of stigma that is the reason scheme isn't popular. I don't
know how to stop it or break it, other than to continue to implement
tools and libraries for a *single* implementation--and make them widely
known, so more people get exposed to it. The more people that get exposed
to something that works, meaning _useful_ and _public_ tools get written in it,
the more people will start to use it and not care about "religion".

As for me? I don't give a rats ass about religion, flamewars, or the
"right way" to do things or any of that shit. I just care if it works,
it is a "reasonable" implementation, and is understandable by the next
person when I leave.

That is all.

-- 
-pete

E-mail address corrupted to stop spam.
Reply mail: psilord at cs dot wisc dot edu
I am responsible for what I say, noone else.
From: Thien-Thi Nguyen
Subject: Re: religion
Date: 
Message-ID: <kk9r8g8yjtw.fsf@glug.org>
Peter Keller <·······@data.upl.cs.wisc.edu> writes:

> As for me? I don't give a rats ass about religion, flamewars, or the
> "right way" to do things or any of that shit. I just care if it works,
> it is a "reasonable" implementation, and is understandable by the next
> person when I leave.

sounds like you should switch to computer engineering.  the life of a computer
scientist is full of all of the above (one would gather from observing these
groups).  on second thought, engineering is the same... never mind, you're
screwed.  (might as well lose your mind now and get it over with.  ;-)

what you'll probably come to see and possibly come to appreciate is that to
find the "reasonable" one has to do all sorts of unreasonable things/thinking.
unless, of course, you have the luxury of being a machine to be programmed by
these pesky humans...  then you can rightly blame them for the mess!

ok, time to step up the medication...

thi
From: MJ Ray
Subject: Re: religion
Date: 
Message-ID: <slrnane3r5.hms.markj+0111@cloaked.freeserve.co.uk>
Peter Keller <·······@data.upl.cs.wisc.edu> wrote:
> [...] Why doesn't oreilly accept books for lisp-like lanauges [...]?

The cynic would say that the amount of money that O'Reilly have sunk into
less powerful languages means that they are maximising the return now.  If
they educated people effectively about the more powerful languages, their
back catalogue sales would go through the floor, wouldn't they?
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240205961130648@naggum.no>
* MJ Ray <··········@cloaked.freeserve.co.uk>
| The cynic would say that the amount of money that O'Reilly have sunk into
| less powerful languages means that they are maximising the return now.  If
| they educated people effectively about the more powerful languages, their
| back catalogue sales would go through the floor, wouldn't they?

  Before the opinionated speculation takes over completely, go look at this:

http://www.oreilly.com/oreilly/author/writeforus_1101.html

  and notice where the word LISP occurs.  Become enlightened.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Fred Gilham
Subject: Re: religion
Date: 
Message-ID: <u7znuwv5k2.fsf@snapdragon.csl.sri.com>
Erik Naggum <····@naggum.no> writes:
>   Before the opinionated speculation takes over completely, go look at this:
> 
> http://www.oreilly.com/oreilly/author/writeforus_1101.html
> 
>   and notice where the word LISP occurs.  Become enlightened.

Well, we can at least be thankful that they didn't put LISP
*immediately* above "pet theories of wombat intercourse".

Though they did put it above "books on topics that have dismal
sales...."

-- 
Fred Gilham ······@csl.sri.com || His word is a creative word, and
when he speaks the good exists as good.  God is neither arbitrary nor
tyrannical.  He is love, and when he expresses his will it is a will
of love.  Hence the good given by God is good for us.-- Jacques Ellul
From: Friedrich Dominicus
Subject: Re: religion
Date: 
Message-ID: <87wuq0prgf.fsf@fbigm.here>
Erik Naggum <····@naggum.no> writes:

> * MJ Ray <··········@cloaked.freeserve.co.uk>
> | The cynic would say that the amount of money that O'Reilly have sunk into
> | less powerful languages means that they are maximising the return now.  If
> | they educated people effectively about the more powerful languages, their
> | back catalogue sales would go through the floor, wouldn't they?
> 
>   Before the opinionated speculation takes over completely, go look at this:
> 
> http://www.oreilly.com/oreilly/author/writeforus_1101.html
> 
>   and notice where the word LISP occurs.  Become enlightened.
This is ridicolous. I can't see any reason for seeing it there. And
fair enough the have books about TeX but refuse some about LaTeX. This
is so idiotic.

Friedrich
From: Christopher Browne
Subject: Re: religion
Date: 
Message-ID: <al7tg2$1o5a9k$3@ID-125932.news.dfncis.de>
Centuries ago, Nostradamus foresaw when Friedrich Dominicus <·····@q-software-solutions.com> would write:
> Erik Naggum <····@naggum.no> writes:
>
>> * MJ Ray <··········@cloaked.freeserve.co.uk>
>> | The cynic would say that the amount of money that O'Reilly have sunk into
>> | less powerful languages means that they are maximising the return now.  If
>> | they educated people effectively about the more powerful languages, their
>> | back catalogue sales would go through the floor, wouldn't they?
>> 
>>   Before the opinionated speculation takes over completely, go look at this:
>> 
>> http://www.oreilly.com/oreilly/author/writeforus_1101.html
>> 
>>   and notice where the word LISP occurs.  Become enlightened.
> This is ridicolous. I can't see any reason for seeing it there. And
> fair enough the have books about TeX but refuse some about LaTeX. This
> is so idiotic.

They haven't accepted books on TeX in some years now.  I don't think
they discriminate against TeX in this any more than they discriminate
against Common Lisp (since "presumably" Scheme, not being named
"LISP," is fair game for books, right?  NOT!).

There are past books in these various areas, but none of it is even
_faintly_ recent.  The Elisp book dates back to 1997.  The TeX book
dates back to 1994.  

The notion that they would have changed policies to reject _further_
submissions on these subjects may be "discriminatory," but I'd not
judge it to be either "ridicolous" or "so idiotic."

(Of course, the author of the book on TeX wrote another book,
published by O'Reilly, that has a section on Scheme, or DSSSL, to be
more precise...  But it isn't a book "on LISP," to be sure, so there's
no particular contradiction with policy in this...)
-- 
(concatenate 'string "cbbrowne" ·@acm.org")
http://www.ntlug.org/~cbbrowne/x.html
Oh,  boy, virtual memory!  Now I'm  gonna make  myself a  really *big*
RAMdisk!
From: Robert Uhl <····@4dv.net>
Subject: Re: religion
Date: 
Message-ID: <m3sn0o4ana.fsf@latakia.dyndns.org>
Christopher Browne <········@acm.org> writes:
> 
> There are past books in these various areas, but none of it is even
> _faintly_ recent.  The Elisp book dates back to 1997.  The TeX book
> dates back to 1994.

I dunno--'97 is fairly recent.  '94 is pushing things a bit though.

It's pretty sad that the market won't bear books on either Lisp or
LaTeX to their satisfaction.

-- 
Robert Uhl <····@4dv.net>
But it's more than that, of course; bad spelling just isn't respectable.
You may, perhaps, want to lament this fact.  You are free to do so.  The
fact remains.                                            --John Mitchell
From: Friedrich Dominicus
Subject: Re: religion
Date: 
Message-ID: <87y9agzb6g.fsf@fbigm.here>
Christopher Browne <········@acm.org> writes:

> 
> The notion that they would have changed policies to reject _further_
> submissions on these subjects may be "discriminatory," but I'd not
> judge it to be either "ridicolous" or "so idiotic."
Well excluding an area in which you usually earn your money (books
about programming and too programming languages or tools) is IMHO
idiotic. What will happen if all people will start asking for Lisp or
LaTeX books will they add an 
- no Java books than?

A reason I can follow is just a very small market. But I'm not sure if
the market is really so small. Well you pointed out other things they
have too books about Emacs and Emacs Lisp which obviously are
Lisps, wo what is their we are not looking after stuff good for?

Friedrich
From: Marco Antoniotti
Subject: Re: religion
Date: 
Message-ID: <y6cfzwo2z3o.fsf@octagon.mrl.nyu.edu>
Friedrich Dominicus <·····@q-software-solutions.com> writes:

> Christopher Browne <········@acm.org> writes:
> 
> > 
> > The notion that they would have changed policies to reject _further_
> > submissions on these subjects may be "discriminatory," but I'd not
> > judge it to be either "ridicolous" or "so idiotic."
> Well excluding an area in which you usually earn your money (books
> about programming and too programming languages or tools) is IMHO
> idiotic. What will happen if all people will start asking for Lisp or
> LaTeX books will they add an 
> - no Java books than?
> 
> A reason I can follow is just a very small market. But I'm not sure if
> the market is really so small. Well you pointed out other things they
> have too books about Emacs and Emacs Lisp which obviously are
> Lisps, wo what is their we are not looking after stuff good for?

I am inclined to accept the "irrationality" hypothesis about the
O'Reilly ban on LISP and LaTeX.

If the market for these books is (and - admittedly - is smaller than
that of "The Idiotitic Guide to SLDJ") then you can just prioritize
them low on the editorial pipeline and devote small resources to them.

But the ban is comprehensive and absolute.  There is no leeway for
market adaptability in there.  Hence I - personally - conclude that it
is an "irrational" choice.  The fact that it does not harm O'Reilly
financially simply makes the overall management indifferent to such
"irrationality".

Cheers

-- 
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group        tel. +1 - 212 - 998 3488
715 Broadway 10th Floor                 fax  +1 - 212 - 995 4122
New York, NY 10003, USA                 http://bioinformatics.cat.nyu.edu
                    "Hello New York! We'll do what we can!"
                           Bill Murray in `Ghostbusters'.
From: Biep @ http://www.biep.org/
Subject: O'Reilly (Was: religion)
Date: 
Message-ID: <al86gu$1ophj5$1@ID-63952.news.dfncis.de>
Marco Antoniotti wrote:
> I am inclined to accept the "irrationality" hypothesis about the
> O'Reilly ban on LISP and LaTeX.
>
> If the market for these books is (and - admittedly - is smaller than
> that of "The Idiotitic Guide to SLDJ") then you can just prioritize
> them low on the editorial pipeline and devote small resources to them.
>
> But the ban is comprehensive and absolute.  There is no leeway for
> market adaptability in there.  Hence I - personally - conclude that it
> is an "irrational" choice.  The fact that it does not harm O'Reilly
> financially simply makes the overall management indifferent to such
> "irrationality".

Well, maybe they have a backlog of books on these subjects, and until the
few resources they do allocate have digested that backlog, they do not
accept any new submissions.

Does anybody know more than the infamous web page tells us?

--
Biep
Reply via any name whatsoever at the main domain corresponding to
http://www.biep.org
From: Christopher Browne
Subject: Re: religion
Date: 
Message-ID: <al85s1$1o8632$1@ID-125932.news.dfncis.de>
A long time ago, in a galaxy far, far away, Friedrich Dominicus <·····@q-software-solutions.com> wrote:
> Christopher Browne <········@acm.org> writes:
>> The notion that they would have changed policies to reject _further_
>> submissions on these subjects may be "discriminatory," but I'd not
>> judge it to be either "ridicolous" or "so idiotic."

> Well excluding an area in which you usually earn your money (books
> about programming and too programming languages or tools) is IMHO
> idiotic. What will happen if all people will start asking for Lisp
> or LaTeX books will they add an - no Java books than?

> A reason I can follow is just a very small market. But I'm not sure
> if the market is really so small. Well you pointed out other things
> they have too books about Emacs and Emacs Lisp which obviously are
> Lisps, wo what is their we are not looking after stuff good for?

Think about TeX, for a moment.  (That's something you'll not likely
feel emotional about.)

Q: Who's the publisher of the "comprehensive/authoritative" works on TeX?

A: The vast majority of the interesting TeX/LaTeX-related books are
published by Addison-Wesley.  That includes the authorities, Knuth and
Lamport, and there's a whole series of other authors that have
produced various "Companion" books.

O'Reilly can't readily compete with that, as they haven't nearly as
authoritative authors.

With Perl, they've got the "top guys" in their stable, and the same is
pretty much true for Python and PHP.  With those language, they _can_
be a "dominant alpha male" in the wolfpack of publishers.  But ramping
up to have even second-string authors about TeX would be a real
challenge.

Looking at TeX, it doesn't make sense for O'Reilly to put in effort to
try to be "dominant," because it would be _VERY_ difficult to unseat
A-W.  The best they can do is to have second- or third-string books,
so they have evidently decided not to even play in the TeX "market."

And much the same holds true for Lisp.  O'Reilly isn't in a position
to "challenge for leadership," so it's not _too_ remarkable for them
to choose to stay out of the game.

With respect to Java, O'Reilly _does_ have a pretty strong stable of
authors, albeit not including the Goslings of the sector.  They may
not have Gosling, but they have a goodly, credible set of authors, and
a strong set of books on a sizable set of Java-related topics.

For Lisp, they haven't any Larry Wall-like authority, and they
apparently figure they haven't a market for a series of Lisp titles
where they could make use of less luminary authors.

Note that in all of this, I haven't once mentioned a single
"technical" issue surrounding the fitness of Lisp.

They may have some issues on that side of things, but I'd think it far
_more_ significant that the editorial staff has members like Larry
Wall that don't have a vastly high regard for Lisp as far as its
present "usefulness."  

There's quite enough "politics," between the "author mix" and the
"editorial preferences" mix, to make it quite pointless to pitch a
Lisp book at O'Reilly.  

Evidently O'Reilly isn't part of the "Lisp Political Party,"
irrespective of whether it is sufficiently inclusive to include Scheme
or Dylan.  And you'll just have to live with that.
-- 
(concatenate 'string "cbbrowne" ·@acm.org")
http://www.ntlug.org/~cbbrowne/
"And  1.1.81 is  officially BugFree(tm),  so  if you  receive any  bug
reports on it, you know they are just evil lies." -- Linus Torvalds
From: Thomas F. Burdick
Subject: Re: religion
Date: 
Message-ID: <xcv7ki0ia69.fsf@hurricane.OCF.Berkeley.EDU>
Christopher Browne <········@acm.org> writes:

> Looking at TeX, it doesn't make sense for O'Reilly to put in effort to
> try to be "dominant," because it would be _VERY_ difficult to unseat
> A-W.  The best they can do is to have second- or third-string books,
> so they have evidently decided not to even play in the TeX "market."
> 
> And much the same holds true for Lisp.  O'Reilly isn't in a position
> to "challenge for leadership," so it's not _too_ remarkable for them
> to choose to stay out of the game.

This line of reasoning works wonderfully ... until you get to Emacs.
I don't know how many times I've had to tell someon more than once to
avoid the O'Reilly Emacs/Elisp books, because they're out of date.
Given that their only competition in this arena is the FSF, I don't
understand why they don't update their still-in-print Emacs books.

-- 
           /|_     .-----------------------.                        
         ,'  .\  / | No to Imperialist war |                        
     ,--'    _,'   | Wage class war!       |                        
    /       /      `-----------------------'                        
   (   -.  |                               
   |     ) |                               
  (`-.  '--.)                              
   `. )----'                               
From: Christopher Browne
Subject: Re: religion
Date: 
Message-ID: <al8b82$1nf2ms$1@ID-125932.news.dfncis.de>
Centuries ago, Nostradamus foresaw when ···@hurricane.OCF.Berkeley.EDU (Thomas F. Burdick) would write:
> Christopher Browne <········@acm.org> writes:
>
>> Looking at TeX, it doesn't make sense for O'Reilly to put in effort to
>> try to be "dominant," because it would be _VERY_ difficult to unseat
>> A-W.  The best they can do is to have second- or third-string books,
>> so they have evidently decided not to even play in the TeX "market."
>> 
>> And much the same holds true for Lisp.  O'Reilly isn't in a position
>> to "challenge for leadership," so it's not _too_ remarkable for them
>> to choose to stay out of the game.
>
> This line of reasoning works wonderfully ... until you get to Emacs.
> I don't know how many times I've had to tell someon more than once to
> avoid the O'Reilly Emacs/Elisp books, because they're out of date.
> Given that their only competition in this arena is the FSF, I don't
> understand why they don't update their still-in-print Emacs books.

Politics can explain that: I don't think Tim O'Reilly likes RMS
terribly much, and they probably _aren't_ foregoing a windfall of
$150K/year out of not selling books about RMS' pet software.

I don't feel the need to expect that people "with money" will
necessarily _not_ do petty things...
-- 
(concatenate 'string "chris" ·@cbbrowne.com")
http://cbbrowne.com/info/finances.html
"Es is nicht gasagt das  es besser wird wenn es  anders wirt.  Wenn es
aber  besser werden soll muss es  anders werden.  (Loosely translated:
Different   is not necessarily  better.   But  better _is_ necessarily
different.)"  -- G. Ch. Lichtenberg
From: Thomas F. Burdick
Subject: Re: religion
Date: 
Message-ID: <xcvelc8go85.fsf@whirlwind.OCF.Berkeley.EDU>
Christopher Browne <········@acm.org> writes:

> Centuries ago, Nostradamus foresaw when ···@hurricane.OCF.Berkeley.EDU (Thomas F. Burdick) would write:
> > Christopher Browne <········@acm.org> writes:
> >
> >> Looking at TeX, it doesn't make sense for O'Reilly to put in effort to
> >> try to be "dominant," because it would be _VERY_ difficult to unseat
> >> A-W.  The best they can do is to have second- or third-string books,
> >> so they have evidently decided not to even play in the TeX "market."
> >> 
> >> And much the same holds true for Lisp.  O'Reilly isn't in a position
> >> to "challenge for leadership," so it's not _too_ remarkable for them
> >> to choose to stay out of the game.
> >
> > This line of reasoning works wonderfully ... until you get to Emacs.
> > I don't know how many times I've had to tell someon more than once to
> > avoid the O'Reilly Emacs/Elisp books, because they're out of date.
> > Given that their only competition in this arena is the FSF, I don't
> > understand why they don't update their still-in-print Emacs books.
> 
> Politics can explain that: I don't think Tim O'Reilly likes RMS
> terribly much, and they probably _aren't_ foregoing a windfall of
> $150K/year out of not selling books about RMS' pet software.

Well, they could certainly focus on XEmacs and its dialect of elisp.
I do wonder how much sales they're missing out on ... I personally
have told about a dozen people so far this year *not* to buy an
O'Reilly book on Emacs, and that's only counting face-to-face
encounters.  Admittedly, I'm a known Emacs user/programmer to these
people, but I'd bet there's a lot of business they're missing.

> I don't feel the need to expect that people "with money" will
> necessarily _not_ do petty things...

I'm guessing that's likely it.

-- 
           /|_     .-----------------------.                        
         ,'  .\  / | No to Imperialist war |                        
     ,--'    _,'   | Wage class war!       |                        
    /       /      `-----------------------'                        
   (   -.  |                               
   |     ) |                               
  (`-.  '--.)                              
   `. )----'                               
From: Neil W. Van Dyke
Subject: Re: religion
Date: 
Message-ID: <jir8g8rt01.fsf@neilvandyke.org>
> > avoid the O'Reilly Emacs/Elisp books, because they're out of date.
> > Given that their only competition in this arena is the FSF, I don't
> > understand why they don't update their still-in-print Emacs books.
> 
> Politics can explain that: I don't think Tim O'Reilly likes RMS
> terribly much, and they probably _aren't_ foregoing a windfall of
> $150K/year out of not selling books about RMS' pet software.

Tim O'Reilly is first and foremost a businessperson.  The FSF's Emacs
and Elisp books are pretty good.  I'm guessing that the market
opportunity for non-FSF Emacs books is mainly limited to people who want
to read about Emacs but don't know that the FSF books exist.  ORA's
resources are probably better invested in more lucrative books.

We also can't shame ORA for not developing a general Scheme book.  There
are already several excellent books established in that space -- and
most are freely-available, to boot.

But note that, historically, ORA hasn't demonstrated a religious
devotion to any particular software platform -- they've gone with the
market.  I'd bet money that they'd go for a Scheme-related book if it
was on, say, a Scheme-based Web framework that lots of people were
actually using and that had need for an ORA book.

(Personally, I'd be perfectly happy if Scheme stayed mostly in academia
and far away from commercial power-plays from Microsoft and the like.)

-- 
                                                        Neil W. Van Dyke
                                             http://www.neilvandyke.org/
From: ozan s yigit
Subject: Re: religion
Date: 
Message-ID: <vi4it1ji0z6.fsf@blue.cs.yorku.ca>
"Neil W. Van Dyke" <····@NOSPAMneilvandyke.org> writes:

> But note that, historically, ORA hasn't demonstrated a religious
> devotion to any particular software platform -- they've gone with the
> market. [...]

that has been my observation since the founding of ORA and the nutshell
series. in any case, i do not understand why we should care about what ORA
does or doesn't publish. there are other publishers that may offer better
home to scheme & lisp books; AW, MIT press, PH, springer etc and some
of the other university presses come to mind...

/where/ is the book that needs a publisher? is it any good?

oz
-- 
it's very difficult to do something small in a meaningful way. -- anon
						(quoted by bruce ross)
From: Sander Vesik
Subject: Re: religion
Date: 
Message-ID: <1031937837.788154@haldjas.folklore.ee>
In comp.lang.scheme Neil W. Van Dyke <····@nospamneilvandyke.org> wrote:
>> > avoid the O'Reilly Emacs/Elisp books, because they're out of date.
>> > Given that their only competition in this arena is the FSF, I don't
>> > understand why they don't update their still-in-print Emacs books.
>> 
>> Politics can explain that: I don't think Tim O'Reilly likes RMS
>> terribly much, and they probably _aren't_ foregoing a windfall of
>> $150K/year out of not selling books about RMS' pet software.
> 
> Tim O'Reilly is first and foremost a businessperson.  The FSF's Emacs
> and Elisp books are pretty good.  I'm guessing that the market
> opportunity for non-FSF Emacs books is mainly limited to people who want
> to read about Emacs but don't know that the FSF books exist.  ORA's
> resources are probably better invested in more lucrative books.
> 
> We also can't shame ORA for not developing a general Scheme book.  There
> are already several excellent books established in that space -- and
> most are freely-available, to boot.
> 
> But note that, historically, ORA hasn't demonstrated a religious
> devotion to any particular software platform -- they've gone with the
> market.  I'd bet money that they'd go for a Scheme-related book if it
> was on, say, a Scheme-based Web framework that lots of people were
> actually using and that had need for an ORA book.

Well, there is say SISC - so if lots of uses that are relevant to "normal"
people using java spring up around SISC that I can image (and would be
willing to bet on) a published by ORA book on it. Think of say the Jython
book.

> 
> (Personally, I'd be perfectly happy if Scheme stayed mostly in academia
> and far away from commercial power-plays from Microsoft and the like.)
> 

-- 
	Sander

+++ Out of cheese error +++
From: Perry E. Metzger
Subject: Re: religion
Date: 
Message-ID: <87y9ag1ler.fsf@snark.piermont.com>
MJ Ray <··········@cloaked.freeserve.co.uk> writes:
> Peter Keller <·······@data.upl.cs.wisc.edu> wrote:
> > [...] Why doesn't oreilly accept books for lisp-like lanauges [...]?
> 
> The cynic would say that the amount of money that O'Reilly have sunk into
> less powerful languages means that they are maximising the return now.  If
> they educated people effectively about the more powerful languages, their
> back catalogue sales would go through the floor, wouldn't they?

The business person would note that since very few people use lisp
they don't want to spend money on something perceived as a niche
market. Books on languages such as Ruby, Perl, Python, etc. all sell
well because lots of people use those languages for their everyday work.

Perry
From: Robert Uhl <····@4dv.net>
Subject: Re: religion
Date: 
Message-ID: <m3ofbc4ah0.fsf@latakia.dyndns.org>
"Perry E. Metzger" <·····@piermont.com> writes:
> 
> The business person would note that since very few people use lisp
> they don't want to spend money on something perceived as a niche
> market.  Books on languages such as Ruby, Perl, Python, etc. all
> sell well because lots of people use those languages for their
> everyday work.

People use Ruby?  Last I looked at it, I wasn't very impressed...

If anything, I wonder if the presence of an O'Reilly book doesn't
_create_ a market to some extent.  If there were an O'Reilly book on
Lisps and another on LaTeX, maybe intelligent languages and
intelligent document preparation might make a comeback:-)

-- 
Robert Uhl <····@4dv.net>
Freedom does not mean freedom just for the things I think I should be
able to do.  Freedom is for all of us.  If people will not speak up
for other people's rights, there will come a day when they will lose
their own.                                           --Tony Lawrence
From: sv0f
Subject: Re: religion
Date: 
Message-ID: <none-0509021345110001@129.59.212.53>
In article <··············@latakia.dyndns.org>, ····@4dv.net (Robert Uhl
<····@4dv.net>) wrote:

>"Perry E. Metzger" <·····@piermont.com> writes:
>>Books on languages such as Ruby, Perl, Python, etc. all
>> sell well because lots of people use those languages for their
>> everyday work.
>
>People use Ruby?  Last I looked at it, I wasn't very impressed...

Exactly my observation.

I can understand why people were excited about Java when it
came along, given the rising distaste for C++ and Microsoft.
I can understand why people found Perl a useful alternative
to a variety of UNIX mini-languages.  I can understand why
people migrated to Python, given that it combines the
systematicity and modernity of real languages with the
convenience of scripted languages.

But what on earth does Ruby have going for it?  People seem
to be chasing the new for new's sake, even though it appears
to offer no single significant improvement.
From: Pascal Costanza
Subject: Re: religion
Date: 
Message-ID: <3D787B3C.B56076FD@cs.uni-bonn.de>
sv0f wrote:
> 

> I can understand why people were excited about Java when it
> came along, given the rising distaste for C++ and Microsoft.
> I can understand why people found Perl a useful alternative
> to a variety of UNIX mini-languages.  I can understand why
> people migrated to Python, given that it combines the
> systematicity and modernity of real languages with the
> convenience of scripted languages.
> 
> But what on earth does Ruby have going for it?  People seem
> to be chasing the new for new's sake, even though it appears
> to offer no single significant improvement.

It's a Smalltalk with more mainstream syntax. Some people like this
combination. A colleague of mine who likes it puts it like this: "It's
object-oriented all the way down like Smalltalk, but the if-statement
still looks like an if-statement."

Some other people I know prefer Ruby over Python because they think that
Python is ill-designed for two reasons: (1) In Python, you alway have to
explicitly carry the "self" reference around. (2) In Python, the naming
conventions for "private" fields is regarded as strange - pre- and
postfixing names with two underscores each - and there is no language
support for enforcing access checks.

Personally, I am not convinced that Ruby's design choices are real
advantages, but it obviously appeals to people who like "pure"
object-orientation.

Pascal

--
Pascal Costanza               University of Bonn
···············@web.de        Institute of Computer Science III
http://www.pascalcostanza.de  R�merstr. 164, D-53117 Bonn (Germany)
From: Perry E. Metzger
Subject: Re: religion
Date: 
Message-ID: <87r8g8gmpk.fsf@snark.piermont.com>
····@4dv.net (Robert Uhl <····@4dv.net>) writes:
> "Perry E. Metzger" <·····@piermont.com> writes:
> > 
> > The business person would note that since very few people use lisp
> > they don't want to spend money on something perceived as a niche
> > market.  Books on languages such as Ruby, Perl, Python, etc. all
> > sell well because lots of people use those languages for their
> > everyday work.
> 
> People use Ruby?  Last I looked at it, I wasn't very impressed...

People use Ruby. Hell, *I* have used Ruby. I know lots of people
adopting it -- as a pure OO language in the smalltalk tradition it has
a lot more interesting things going for it than, say, Perl. Also, it
is trivial to get a lot of work done in it quickly -- as with Perl and
other similar languages it has a lot of API bindings for things you
need to call already written for you. Ruby also has a very friendly
and active community -- they don't get flames on their lists the way
they happen around here. As a result of all this, the community is
growing way fast.


-- 
Perry E. Metzger		·····@piermont.com
--
"Ask not what your country can force other people to do for you..."
From: Sander Vesik
Subject: Re: religion
Date: 
Message-ID: <1031261081.171544@haldjas.folklore.ee>
In comp.lang.scheme Robert Uhl <····@4dv.net> <····@4dv.net> wrote:
> "Perry E. Metzger" <·····@piermont.com> writes:
>> 
>> The business person would note that since very few people use lisp
>> they don't want to spend money on something perceived as a niche
>> market.  Books on languages such as Ruby, Perl, Python, etc. all
>> sell well because lots of people use those languages for their
>> everyday work.
> 
> People use Ruby?  Last I looked at it, I wasn't very impressed...

Yes, lots of people use ruby. Maybe not for anything practical
but its a cool language that is 'in' and lots of peopel want to 
know it and hence buy books about it.

> 
> If anything, I wonder if the presence of an O'Reilly book doesn't
> _create_ a market to some extent.  If there were an O'Reilly book on
> Lisps and another on LaTeX, maybe intelligent languages and
> intelligent document preparation might make a comeback:-)
> 

No, they would need to be popular first.

-- 
	Sander

+++ Out of cheese error +++
From: Friedrich Dominicus
Subject: Re: religion
Date: 
Message-ID: <87lm6fg1wj.fsf@fbigm.here>
Sander Vesik <······@haldjas.folklore.ee> writes:

> In comp.lang.scheme Robert Uhl <····@4dv.net> <····@4dv.net> wrote:
> > "Perry E. Metzger" <·····@piermont.com> writes:
> >> 
> >> The business person would note that since very few people use lisp
> >> they don't want to spend money on something perceived as a niche
> >> market.  Books on languages such as Ruby, Perl, Python, etc. all
> >> sell well because lots of people use those languages for their
> >> everyday work.
> > 
> > People use Ruby?  Last I looked at it, I wasn't very impressed...
> 
> Yes, lots of people use ruby. Maybe not for anything practical
Do you feel this is a good idea to post? 

If people are using Ruby why should they use it of inpractial things?
The use it because it's usefule to them and the surely do not just
play around with it but do the work they have to do so Ruby is very 
*practical* for them. As is Common Lisp for us.

Friedrich
From: Sander Vesik
Subject: Re: religion
Date: 
Message-ID: <1031938312.869676@haldjas.folklore.ee>
In comp.lang.scheme Friedrich Dominicus <·····@q-software-solutions.com> wrote:
> Sander Vesik <······@haldjas.folklore.ee> writes:
> 
>> In comp.lang.scheme Robert Uhl <····@4dv.net> <····@4dv.net> wrote:
>> > "Perry E. Metzger" <·····@piermont.com> writes:
>> >> 
>> >> The business person would note that since very few people use lisp
>> >> they don't want to spend money on something perceived as a niche
>> >> market.  Books on languages such as Ruby, Perl, Python, etc. all
>> >> sell well because lots of people use those languages for their
>> >> everyday work.
>> > 
>> > People use Ruby?  Last I looked at it, I wasn't very impressed...
>> 
>> Yes, lots of people use ruby. Maybe not for anything practical
> Do you feel this is a good idea to post? 
> 
> If people are using Ruby why should they use it of inpractial things?
> The use it because it's usefule to them and the surely do not just
> play around with it but do the work they have to do so Ruby is very 
> *practical* for them. As is Common Lisp for us.

Lots of people do a lot of coding that is not done for the purpose of
it giving a practical result or even because its part of their work.
All python code I have written so far is such for example - I just
wanted to learn python. Writing code in a language does not have to
have any other purpose than explorartion of that language and learning.

> 
> Friedrich

-- 
	Sander

+++ Out of cheese error +++
From: Friedrich Dominicus
Subject: Re: religion
Date: 
Message-ID: <87k7m1q7eg.fsf@fbigm.here>
Peter Keller <·······@data.upl.cs.wisc.edu> writes:

> In comp.lang.scheme Christopher Browne <········@acm.org> wrote:
> : In the last exciting episode, Peter Keller <·······@data.upl.cs.wisc.edu> wrote::
> :> I'm beginning to believe that the reason scheme isn't as popular as
> :> the rest of the languages isn't so much as there are many
> :> implementations, but more so that no-one understands how to market
> :> it worth a damn.
> 
> :> Seen any oreilly books on scheme, or scheme interfaces to SQL or
> :> anything like that? My point exactly.
> 
> : O'Reilly also doesn't have books on Common Lisp, and aren't accepting
> : new titles on TeX or _any_ form of Lisp, so I'd not count that as
> : being of much interest.
> 
> Well, I can buy an oreilly book on "mastering regular expressions" but not
> "mastering macros". Why? Why doesn't oreilly accept books for lisp-like
> lanauges, and why isn't there a book out there like a "macro cookbook"
> which talks about the various forms of macro representation and what
> you can do with them with good *and relevant* examples by any publisher
> at all?

what's wrong with OnLisp?
http://www.paulgraham.com/onlisptext.html

It does exactly what you ask for explaining Macros. 

Regards
Friedrich
From: Peter Keller
Subject: Re: religion
Date: 
Message-ID: <3d76fbd0$0$1125$80265adb@spool.cs.wisc.edu>
In comp.lang.scheme Friedrich Dominicus <·····@q-software-solutions.com> wrote:
: what's wrong with OnLisp?
: http://www.paulgraham.com/onlisptext.html

: It does exactly what you ask for explaining Macros. 

Hmm.... nothing is wrong with it, other than I didn't know about it. :)
I usually only code in the scheme domains, not full common lisp.

Does it describe R5RS macros too? Or can the techniques in this book be 
readily translated to R5RS macros as well?

-- 
-pete

E-mail address corrupted to stop spam.
Reply mail: psilord at cs dot wisc dot edu
I am responsible for what I say, noone else.
From: Friedrich Dominicus
Subject: Re: religion
Date: 
Message-ID: <871y88r66g.fsf@fbigm.here>
Peter Keller <·······@data.upl.cs.wisc.edu> writes:

> In comp.lang.scheme Friedrich Dominicus <·····@q-software-solutions.com> wrote:
> : what's wrong with OnLisp?
> : http://www.paulgraham.com/onlisptext.html
> 
> : It does exactly what you ask for explaining Macros. 
> 
> Hmm.... nothing is wrong with it, other than I didn't know about it. :)
> I usually only code in the scheme domains, not full common lisp.
> 
> Does it describe R5RS macros too? 
No.

> Or can the techniques in this book be 
> readily translated to R5RS macros as well?
I don't think so, probably just partly. AFAIK do Schemes offer usually
more than one macro system there are the R5RS Macros but there are
macro systems that work simular to Common Lisps defmacro. But that's
probably a questoin for c.l.scheme

Regards
Friedrich
From: Sander Vesik
Subject: Re: religion
Date: 
Message-ID: <1031226505.395176@haldjas.folklore.ee>
In comp.lang.scheme Peter Keller <·······@data.upl.cs.wisc.edu> wrote:
> In comp.lang.scheme Christopher Browne <········@acm.org> wrote:
> : In the last exciting episode, Peter Keller <·······@data.upl.cs.wisc.edu> wrote::
> :> I'm beginning to believe that the reason scheme isn't as popular as
> :> the rest of the languages isn't so much as there are many
> :> implementations, but more so that no-one understands how to market
> :> it worth a damn.
> 
> :> Seen any oreilly books on scheme, or scheme interfaces to SQL or
> :> anything like that? My point exactly.
> 
> : O'Reilly also doesn't have books on Common Lisp, and aren't accepting
> : new titles on TeX or _any_ form of Lisp, so I'd not count that as
> : being of much interest.
> 
> Well, I can buy an oreilly book on "mastering regular expressions" but not
> "mastering macros". Why? Why doesn't oreilly accept books for lisp-like
> lanauges, and why isn't there a book out there like a "macro cookbook"
> which talks about the various forms of macro representation and what
> you can do with them with good *and relevant* examples by any publisher
> at all? The regex book pulled it off and there are a BUNCH of regular
> expression styles out there that are incompatible.

O'Reilly is a for-profit comapny last I checked so whetever or not they
will acceptys omething will very probably depend on whetever they see a
market for it. If I was them, I would not accept a scheme or lisp book
either.

> 
> : And the following bit of Guile code does _exactly_ what it would be
> : expected to do, namely interface with PostgreSQL, and pull a set of
> : entries.
> 
> I wasn't talking about implementations. I was talking about why I can
> wander to a book store and see shelf upon shelf of ASP/JAVA/PERL/C++/SQL
> books and then find exactly one grungy SICP book mislocated into the
> COBOL section--the only scheme book to be found.
> 

Because the number of at large scheme developers strating new projects
or releasing new versions of existing ones is terribly slow if you are
interested in the 'per day' amounts. Which also means per-day number
of peopel thinking of participating in such (or starting their own) is
also terribly slow => no books.

> 
> That is all.
> 

-- 
	Sander

+++ Out of cheese error +++
From: Duane Rettig
Subject: Re: religion
Date: 
Message-ID: <4u1l4778h.fsf@beta.franz.com>
Sander Vesik <······@haldjas.folklore.ee> writes:

> O'Reilly is a for-profit comapny last I checked so whetever or not they
> will acceptys omething will very probably depend on whetever they see a
> market for it. If I was them, I would not accept a scheme or lisp book
> either.

This is true for all for-profit companies, even Lisp companies.  At
various times, we've also considered publishing hard-bound books
and rejected the idea, because the loss-leader was more expensive
than if we were to put our marketing budget into something with more
return-on-investment.

But let's back up a bit.  This portion of this thread has exploded
because many people have taken one person's definition of "vibrancy"
in a language as fact.  What we really should be doing is examining
the _reason_ why Lisp books are not profitable.  I see it in
precisely the opposite way; the proliferation of books on a language
are a sign of that languuage's weakness, not its strength.

Besides profit (or, at least, minimizing loss) there are other reasons
why potential publishers will put out a book.  If it fills a hole
in the subject area, it might be worth putting the book out anyway.
But usually, only persons or companies that are interested in the
subject material will be the ones to do this, and only if the holes
are perceived as large enough.

This is a classic case of "worse is better".  The languages with
the largest holes get the most proliferation of books (and libraries,
for that matter).  Do I think that Lisp is perfect?  Not at all.
Do I think Lisp can use more books in its repertoire?  Absolutely.
But I do think that the original poster, charging in and talking
about vibrancy and arrogance, arrogantly got it backward.

-- 
Duane Rettig    ·····@franz.com    Franz Inc.  http://www.franz.com/
555 12th St., Suite 1450               http://www.555citycenter.com/
Oakland, Ca. 94607        Phone: (510) 452-2000; Fax: (510) 452-0182   
From: Perry E. Metzger
Subject: Re: religion
Date: 
Message-ID: <87n0qwglum.fsf@snark.piermont.com>
Duane Rettig <·····@franz.com> writes:
> This is a classic case of "worse is better".  The languages with
> the largest holes get the most proliferation of books (and libraries,
> for that matter).  Do I think that Lisp is perfect?  Not at all.
> Do I think Lisp can use more books in its repertoire?  Absolutely.
> But I do think that the original poster, charging in and talking
> about vibrancy and arrogance, arrogantly got it backward.

Starting from the top:

1) I think Lisp (and when I say lisp I mean CL, Scheme, etc. -- I'm
not bigoted) is the tops. There's nothing quite like it. I wrote my
first lisp program a very long time ago (the computer in question was
36 bits if that says anything to you), fell in love with it and have
always dreamed of using lisp day to day for all my work since. I
happen to lean in the direction of Scheme for various reasons but I
don't see any reason to claim the CL people are horrible or what have
you. I try to like everyone -- even people who don't use lisp.

2) What I'm pointing out is that, believe it or not, the lisp
community is very small, and although the absolute numbers of
lispheads keeps growing I think that in terms of the percentage of the
world's software development being done in lisp the language is not
exactly winning.

Now one can take this in a couple of different ways. One can say "no,
evil person, go away, do not say this." Or one can say "gee, maybe he
has a point. Maybe there really are more Perl users than lisp users,
and maybe it would be interesting to figure out why. Maybe there is a
lot more in the way of tools for the Java crowd than for the Lisp
crowd. Maybe it would be interesting to figure out ways to improve
that."

Having commercial books written about your language isn't a sign that
you're "better" in some moral sense, but it is an indication that
there are enough users to drive a secondary market for documentation,
software tools, etc. "Market" doesn't even necessarily mean money --
just that there is enough use that people start building and writing
the stuff that makes the language easier to use.

Does this mean I think it would be good to do dumb things to try to
increase "popularity"? No. The issue isn't the lack of infix notation
or whatever else. However, it is sort of a bad sign that a lot of the
energy in the community is used on fratricide instead of "hey, look a
this! a cool rendering package I wrote! anyone want it?" which is what
you find in a lot of the other programming communities.


-- 
Perry E. Metzger		·····@piermont.com
--
"Ask not what your country can force other people to do for you..."
From: Fred Gilham
Subject: Re: religion
Date: 
Message-ID: <u7d6rst0qk.fsf@snapdragon.csl.sri.com>
> 2) What I'm pointing out is that, believe it or not, the lisp
> community is very small, and although the absolute numbers of
> lispheads keeps growing I think that in terms of the percentage of
> the world's software development being done in lisp the language is
> not exactly winning.

One thing to keep in mind is that, when the population of a community
grows, it tends to `back up the bell curve'.  A small community can be
comprised of mostly exceptional individuals but a larger community
runs out.

So `winning' in the context of `community' might simply mean
continuing to attract people from the right end of the bell curve.

-- 
Fred Gilham                     ······@csl.sri.com
We have yet to find the Galileo who will question
our me-centred universe. --- Christina Odone
From: Craig Brozefsky
Subject: Re: religion
Date: 
Message-ID: <87ofbbsqoc.fsf@piracy.red-bean.com>
"Perry E. Metzger" <·····@piermont.com> writes:
> 2) What I'm pointing out is that, believe it or not, the lisp
> community is very small, and although the absolute numbers of
> lispheads keeps growing I think that in terms of the percentage of
> the world's software development being done in lisp the language is
> not exactly winning.

Scores of people have posted the same question, or observation, or
advice, or comment, or complaint, or whatever it was registered as at
the time, in the last year alone.  Noone has any illusions about what
percentage of the body of programmers in the world are using CL or
scheme.  In fact this scenario plays it out so goddamn often in this
newsgroup it can be considered a continuous troll/thread.  And you
wonder why people are getting upset with the next thumb-twiddling
kibutzer that continues it?

> Does this mean I think it would be good to do dumb things to try to
> increase "popularity"? No. The issue isn't the lack of infix notation
> or whatever else. However, it is sort of a bad sign that a lot of the
> energy in the community is used on fratricide instead of "hey, look a
> this! a cool rendering package I wrote! anyone want it?" which is what
> you find in a lot of the other programming communities.

This is usenet, and the goings on here are hardly a valid indicator of
what the lisp community at large is doing.  In fact, Usenet is perhaps
one of the worse indicators, since it has a very strong idiot
attractor and as a medium unto itself, an extremely low
signal-to-noise ratio.

I say, put up or shut up.

-- 
Sincerely yours in arrogant lisp jackassery,
Craig Brozefsky <·····@red-bean.com>  
Free Scheme/Lisp Software  http://www.red-bean.com/~craig
From: Paolo Amoroso
Subject: Re: religion
Date: 
Message-ID: <cZB4PVRpGPTxX6ZYuOw0C3lsP5Oq@4ax.com>
[Followupo posted to comp.lang.lisp only]

On 05 Sep 2002 17:36:33 -0400, "Perry E. Metzger" <·····@piermont.com>
wrote:

> 2) What I'm pointing out is that, believe it or not, the lisp
> community is very small, and although the absolute numbers of
> lispheads keeps growing I think that in terms of the percentage of the
> world's software development being done in lisp the language is not
> exactly winning.
> 
> Now one can take this in a couple of different ways. One can say "no,
> evil person, go away, do not say this." Or one can say "gee, maybe he

Another option: try to learn more about the actual Lisp community and the
language's use in industry and elsewhere. Even well known experts such as
Peter Norvig may not keep up to date with what goes on in the community
(see a recent comp.lang.lisp thread about this). You seemed to be unaware
of cCLan.


> has a point. Maybe there really are more Perl users than lisp users,
> and maybe it would be interesting to figure out why. Maybe there is a

If you wrote your first Lisp programs on a 36-bit machine, you should
already know at least part of the answer (hint: check a well known paper by
Richard Gabriel).


> or whatever else. However, it is sort of a bad sign that a lot of the
> energy in the community is used on fratricide instead of "hey, look a
> this! a cool rendering package I wrote! anyone want it?" which is what
> you find in a lot of the other programming communities.

May we have a look at your own rendering package?


Paolo
-- 
EncyCMUCLopedia * Extensive collection of CMU Common Lisp documentation
http://www.paoloamoroso.it/ency/README
From: Christopher Browne
Subject: Re: religion
Date: 
Message-ID: <al7tg1$1o5a9k$1@ID-125932.news.dfncis.de>
The world rejoiced as Peter Keller <·······@data.upl.cs.wisc.edu> wrote:
> In comp.lang.scheme Christopher Browne <········@acm.org> wrote:
> : In the last exciting episode, Peter Keller <·······@data.upl.cs.wisc.edu> wrote::
> :> I'm beginning to believe that the reason scheme isn't as popular as
> :> the rest of the languages isn't so much as there are many
> :> implementations, but more so that no-one understands how to market
> :> it worth a damn.
>
> :> Seen any oreilly books on scheme, or scheme interfaces to SQL or
> :> anything like that? My point exactly.
>
> : O'Reilly also doesn't have books on Common Lisp, and aren't accepting
> : new titles on TeX or _any_ form of Lisp, so I'd not count that as
> : being of much interest.
>
> Well, I can buy an oreilly book on "mastering regular expressions"
> but not "mastering macros". Why? Why doesn't oreilly accept books
> for lisp-like lanauges, and why isn't there a book out there like a
> "macro cookbook" which talks about the various forms of macro
> representation and what you can do with them with good *and
> relevant* examples by any publisher at all? The regex book pulled it
> off and there are a BUNCH of regular expression styles out there
> that are incompatible.

I do not know.  I _do_ know that they specifically disqualify books on
Lisp and LaTeX.

<http://www.oreilly.com/oreilly/author/writeforus_1101.html>

I'm quite sure that irrespective of any disputing that might take
place around here about whether Scheme is or is not "a Lisp," it is
doubtless enough _like_ Lisp to qualify as such for _their_ purposes.
-- 
(reverse (concatenate 'string ·············@" "enworbbc"))
http://www.ntlug.org/~cbbrowne/lisp.html
((lambda (foo) (bar foo)) (baz))
From: Eduardo Muñoz
Subject: Re: religion
Date: 
Message-ID: <u7ki0tl2m.fsf@jet.es>
Peter Keller <·······@data.upl.cs.wisc.edu> writes:

> In comp.lang.scheme Christopher Browne <········@acm.org> wrote:
> : In the last exciting episode, Peter Keller <·······@data.upl.cs.wisc.edu> wrote::
> :> I'm beginning to believe that the reason scheme isn't as popular as
> :> the rest of the languages isn't so much as there are many
> :> implementations, but more so that no-one understands how to market
> :> it worth a damn.
> 
> :> Seen any oreilly books on scheme, or scheme interfaces to SQL or
> :> anything like that? My point exactly.
> 
> : O'Reilly also doesn't have books on Common Lisp, and aren't accepting
> : new titles on TeX or _any_ form of Lisp, so I'd not count that as
> : being of much interest.
> 
> Well, I can buy an oreilly book on "mastering regular expressions" but not
> "mastering macros". Why? Why doesn't oreilly accept books for lisp-like
> lanauges, and why isn't there a book out there like a "macro cookbook"
> which talks about the various forms of macro representation and what
> you can do with them with good *and relevant* examples by any publisher
> at all? The regex book pulled it off and there are a BUNCH of regular
> expression styles out there that are incompatible.

IMHO the "problem" for O'Reilly is that TeX and
Lisp are stable technologies while Perl, Java,
regexps and the like are everchanging things that
allow then to sell almost the same book every year
or two.

This is the holy grail of IT now. A constant
stream of revenue based on selling little more
than smoke and mirrors.


-- 

Eduardo Mu�oz
From: Marco Antoniotti
Subject: Re: religion
Date: 
Message-ID: <y6c8z2g4ijr.fsf@octagon.mrl.nyu.edu>
Peter Keller <·······@data.upl.cs.wisc.edu> writes:

        ...
> 
> I'm beginning to believe that the reason scheme isn't as popular as the
> rest of the languages isn't so much as there are many implementations, but
> more so that no-one understands how to market it worth a damn.

When the spec is small, whipping up another implementation is (too)
easy.

> Seen any oreilly books on scheme, or scheme interfaces to SQL or anything like
> that? My point exactly.

You have not seen a book on Common Lisp (or Scheme) in the O'Reilly
lineup because the have a ban on "Lisp" and "LaTeX".

        http://www.oreilly.com/oreilly/author/writeforus_1101.html

Cheers

-- 
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group        tel. +1 - 212 - 998 3488
715 Broadway 10th Floor                 fax  +1 - 212 - 995 4122
New York, NY 10003, USA                 http://bioinformatics.cat.nyu.edu
                    "Hello New York! We'll do what we can!"
                           Bill Murray in `Ghostbusters'.
From: Peter Keller
Subject: Re: religion
Date: 
Message-ID: <3d777601$0$1126$80265adb@spool.cs.wisc.edu>
In comp.lang.scheme Marco Antoniotti <·······@cs.nyu.edu> wrote:

: Peter Keller <·······@data.upl.cs.wisc.edu> writes:

: You have not seen a book on Common Lisp (or Scheme) in the O'Reilly
: lineup because the have a ban on "Lisp" and "LaTeX".

:         http://www.oreilly.com/oreilly/author/writeforus_1101.html

Huh. I didn't know that....

-- 
-pete

E-mail address corrupted to stop spam.
Reply mail: psilord at cs dot wisc dot edu
I am responsible for what I say, noone else.
From: Marco Antoniotti
Subject: Re: religion
Date: 
Message-ID: <y6cwuq02zqc.fsf@octagon.mrl.nyu.edu>
Peter Keller <·······@data.upl.cs.wisc.edu> writes:

> In comp.lang.scheme Marco Antoniotti <·······@cs.nyu.edu> wrote:
> 
> : Peter Keller <·······@data.upl.cs.wisc.edu> writes:
> 
> : You have not seen a book on Common Lisp (or Scheme) in the O'Reilly
> : lineup because the have a ban on "Lisp" and "LaTeX".
> 
> :         http://www.oreilly.com/oreilly/author/writeforus_1101.html
> 
> Huh. I didn't know that....

Neither did I.  I was very surprised to see that.  Somebody at
O'Reilly must really hate us. :)  (And Leslie Lamport as well :) )

Cheers

-- 
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group        tel. +1 - 212 - 998 3488
715 Broadway 10th Floor                 fax  +1 - 212 - 995 4122
New York, NY 10003, USA                 http://bioinformatics.cat.nyu.edu
                    "Hello New York! We'll do what we can!"
                           Bill Murray in `Ghostbusters'.
From: Kenny Tilton
Subject: Re: religion
Date: 
Message-ID: <3D779089.8060401@nyc.rr.com>
Marco Antoniotti wrote:
>   Somebody at
> O'Reilly must really hate us. :)

The good news: we have a metric to tell us when Lisp will have returned 
from exile to reclaim its rightful place: "Did we say 'no Lisp'?"

Now who will volunteer to check that web page every day?

kenny
From: Takehiko Abe
Subject: Re: religion
Date: 
Message-ID: <keke-0609020119270001@solg4.keke.org>
In article <···············@octagon.mrl.nyu.edu>, 
Marco Antoniotti wrote:

> Neither did I.  I was very surprised to see that.  Somebody at
> O'Reilly must really hate us. :)  (And Leslie Lamport as well :) )
> 

Love/Hate is mutual. Does Lisp community love O'Reilly? 
I doubt that. 

**

I hate O'Reilly books. It is puzzling that those who demand all
software to be free and chant 'Proprietary!' in every occasion
are so willing to pay for O'Reilly's books.

-- 
This message was not sent to you unsolicited.
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240235922680005@naggum.no>
* ····@ma.ccom (Takehiko Abe)
| Love/Hate is mutual. Does Lisp community love O'Reilly?  I doubt that.

  The sad thing about love and hate is that they are both normally unrequited.
  Some have argued that all real love is inrequited.  What makes this so sad,
  is that those who hate somebody else who don't give a flying fsck about them
  is that they make a lot of noise just to be heard and seen as human beings.
  Invisibility in the eyes of somebody you love or hate can be really painful
  and the reason people do so many stupid things when these emotions place
  themselves between their intellect and their actions is precisely that they
  crave attention, much like a lonely cat.

| I hate O'Reilly books. It is puzzling that those who demand all software to
| be free and chant 'Proprietary!' in every occasion are so willing to pay for
| O'Reilly's books.

  It seems to be a political decision on your part.  I like their books because
  they usually manage to find authors who are way smarter than the average
  crop many other publishers seem to attract.  Their Nutshell Handbooks are
  quite frequently /the/ essential reference.  For the areas where I have no
  desire to become an expert, I have also found their offerings very good.

  There is no doubt that O'Reilly have profited on the Open Source movement.
  I can hardly fault them for that -- I have some 40 volumes from them to date
  and they have even sent me a couple books free of charge, so I can only say
  that I am quite satisfied with them.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Takehiko Abe
Subject: Re: religion
Date: 
Message-ID: <keke-0609022156290001@solg4.keke.org>
In article <················@naggum.no>, Erik Naggum wrote:

> * ····@ma.ccom (Takehiko Abe)
> | Love/Hate is mutual. Does Lisp community love O'Reilly?  I doubt that.
> 
>   The sad thing about love and hate is that they are both normally unrequited.
>   Some have argued that all real love is inrequited.

I'm not familiar with the words 'requited' and 'inrequited'. but I think
I can safely say that love is more likely to induce love on the other end,
whereas hate induces hate. I've also heard the argument that real love does
not demand anything on return, and I agree.  Loving somebody/something is
rewarding in itself.

>   What makes this so sad,
>   is that those who hate somebody else who don't give a flying fsck about them
>   is that they make a lot of noise just to be heard and seen as human beings.
>   Invisibility in the eyes of somebody you love or hate can be really painful
>   [...]

If I'm invisible to someone, I think I will quickly lose my feeling toward
them too if I had any in the first place. It's called 'rapport' or something?
But I guess I'm talking about entirely different matter from what you wrote,
and I'm beginning to feel that I don't know what I'm talking about, so
I should quit.

> 
> | I hate O'Reilly books. It is puzzling that those who demand all software to
> | be free and chant 'Proprietary!' in every occasion are so willing to pay for
> | O'Reilly's books.
> 
>   It seems to be a political decision on your part. 

Yes. I must admit that. And my saying that I hate O'Reilly books was not
based on solid examination of their products. Now that you and Friedrich
Dominicus say you like and value their books, I will try revising my
opinion.

btw, Picking up a CS book here in Osaka (japan) is hard. It is especially
so after the arrival of Amazon.com, which has caused local bookstores to
cut the space for their imported books section significantly.

regards,
Abe

-- 
This message was not sent to you unsolicited.
From: Nils Goesche
Subject: Re: religion
Date: 
Message-ID: <lkofbbckxl.fsf@pc022.bln.elmeg.de>
····@ma.ccom (Takehiko Abe) writes:

> btw, Picking up a CS book here in Osaka (japan) is hard.

No Kuni-Kuniya or what's-it-called there? :-) I think you can also buy
over the internet from them, even to foreign countries.

Regards,
-- 
Nils Goesche
"Don't ask for whom the <CTRL-G> tolls."

PGP key ID 0x0655CFA0
From: Takehiko Abe
Subject: Re: religion
Date: 
Message-ID: <keke-0609022357190001@solg4.keke.org>
In article <··············@pc022.bln.elmeg.de>, Nils Goesche wrote:

> > btw, Picking up a CS book here in Osaka (japan) is hard.
> 
> No Kuni-Kuniya or what's-it-called there? :-) 

That's Kino-Kuni-Ya. Yes, we have them here too. Also there
are some other bookstores as large. 

> I think you can also buy over the internet from them, even to
> foreign countries.

Yes. But the problem is that I cannot discover new books
nor throughly examined them before I pay.  As I said, they
all reduced the space for imported books significantly,
because, I guess, people started to buy from Amazon who put
much lower price tags on the same books.

regards,
abe

-- 
This message was not sent to you unsolicited.
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240322403294208@naggum.no>
* ····@ma.ccom (Takehiko Abe)
| I'm not familiar with the words 'requited' and 'inrequited'.

  A dozen excellent, /free/ dictionaries on the Net and you cannot be bothered
  to look things up?  And when you quote a word, you do not even have editor
  support for proper cut and paste to spell it the same way it was used?  This
  is, I guess the ultimate example of how unrequited Usenet contributions are.
  It is almost amusingly so.

| but I think I can safely say that love is more likely to induce love on the
| other end, whereas hate induces hate.

  Hm.  I find this rather amazing, but it far off-topic.

| I've also heard the argument that real love does not demand anything on
| return, and I agree.  Loving somebody/something is rewarding in itself.

  You /really/ need to look up "unrequited".  May I suggest m-w.com?

| btw, Picking up a CS book here in Osaka (japan) is hard. It is especially so
| after the arrival of Amazon.com, which has caused local bookstores to cut
| the space for their imported books section significantly.

  Interesting.  My local bookstores are also playing it safe these days.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Takehiko Abe
Subject: Re: religion
Date: 
Message-ID: <keke-0709021303360001@solg4.keke.org>
In article <················@naggum.no>, Erik Naggum wrote:

> | I'm not familiar with the words 'requited' and 'inrequited'.
> 
>   A dozen excellent, /free/ dictionaries on the Net and you cannot be bothered
>   to look things up? 

I did. Does "I'm not familiar with the words..." have to mean "I don't
know the words and I'm too dumb to consult a dictionary."? No other a
bit forgiving interpretation possible?

I looked it up the the word, but since it was still new to me, I was
not confident whether I understood your use of the words correctly.

>   And when you quote a word, you do not even have editor
>   support for proper cut and paste to spell it the same way it was used? 

The first word I looked up was 'requite' and I was aware that you did
not write 'requited'. Oops. Sorry, I realized that I misspelled doubly.

>   This is, I guess the ultimate example of how unrequited Usenet
>   contributions are. It is almost amusingly so.

Please blame Usenet.

> 
> | I've also heard the argument that real love does not demand anything on
> | return, and I agree.  Loving somebody/something is rewarding in itself.
> 
>   You /really/ need to look up "unrequited".  May I suggest m-w.com?

which I did. My dictionary says 'requite' means "give sth in return
for sth else" with the example: "Will she ever requite my love?", and
"not returned or rewarded" for 'unrequited'. I meant loving somebody
is /rewarding/ even if you were confined in a jail cell and have no
contact with outside world whatsoever. or something like that.

-- 
This message was not sent to you unsolicited.
From: Friedrich Dominicus
Subject: Re: religion
Date: 
Message-ID: <87heh3g1nb.fsf@fbigm.here>
····@ma.ccom (Takehiko Abe) writes:

> In article <···············@octagon.mrl.nyu.edu>, 
> Marco Antoniotti wrote:
> 
> > Neither did I.  I was very surprised to see that.  Somebody at
> > O'Reilly must really hate us. :)  (And Leslie Lamport as well :) )
> > 
> 
> Love/Hate is mutual. Does Lisp community love O'Reilly? 
> I doubt that. 
Love/hate are a bit too strong feelings IMHO. Anyway I can assure you
that I like Common Lisp and I do like O'Reilly books too. They both
share some things IMHO. E.g they are both above average. Willl say
Lisp has some ideas in it which makes it outstanding and most of the
O'Reilly books are outstanding too. 

Well not all of course but you always can find a hair in the soup
(don't know if that is good transliteration from German)

Regards
Friedrich
From: Erik Naggum
Subject: Re: religion
Date: 
Message-ID: <3240187078453645@naggum.no>
* Perry E. Metzger
| When I asked one of the high cognoscs about the absence of an i/o
| multiplexing primitive in his dialect because I needed one for doing some
| event driven programming, I was regaled with why event driven programming is
| stupid and ugly and I should be using threads instead.  I've had a dozen
| such experiences of late.

  And this is an argument for what, precisely?  Like, /whose/ arrogance?

| I think calling anyone an "arrogant little snot" tends to lessen the power
| of one's claim to humility, don't you?

  No, of course not.  I am humble in the face of the field of mathematics, in
  awe of the great number of geniuses before me, greatly inspired by the works
  of many brilliant minds that could both conceive of the most elegant
  concepts and their intricate interrelationships and formulate them so
  cleanly and briefly in books that I could read and grasp their ideas in but
  a minute fraction of the time it took them to develop them, and yet I am
  disgusted to the deepness of my soul by the lack of even the most basic
  arithmetical skills and the rampant innumeracy of many journalists and
  politicians, because those are people who are so unequivocally /not/ humble
  in the face of mathematics, lacking any and all appreciation of the field,
  yet have the gall to abuse simple results in the field in a way that shows
  an utter disrespect for all the great minds that made it possible for them
  to have a chance to grasp mathematical ideas in their compulsory education,
  but discarded that chance and instead spit in the face of every mathematical
  thinker with their every breath.

| Ah, that explains it. Thank you for enlightening me.

  You're welcome.  I note that the intelligent /exchange/ of ideas has ceased.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.
From: Thomas Stegen
Subject: Re: religion
Date: 
Message-ID: <3D77BB14.3050504@cis.strath.ac.uk>
Erik Naggum wrote:

>   No, of course not.  I am humble in the face of the field of mathematics, in
>   awe of the great number of geniuses before me, greatly inspired by the works
>   of many brilliant minds that could both conceive of the most elegant
>   concepts and their intricate interrelationships and formulate them so
>   cleanly and briefly in books that I could read and grasp their ideas in but
>   a minute fraction of the time it took them to develop them, and yet I am
>   disgusted to the deepness of my soul by the lack of even the most basic
>   arithmetical skills and the rampant innumeracy of many journalists and
>   politicians, because those are people who are so unequivocally /not/ humble
>   in the face of mathematics, lacking any and all appreciation of the field,
>   yet have the gall to abuse simple results in the field in a way that shows
>   an utter disrespect for all the great minds that made it possible for them
>   to have a chance to grasp mathematical ideas in their compulsory education,
>   but discarded that chance and instead spit in the face of every mathematical
>   thinker with their every breath.
> 


Hear hear!

-- 
Thomas Stegen.
From: Tim Bradshaw
Subject: Re: religion
Date: 
Message-ID: <ey3u1l4dema.fsf@cley.com>
* Perry E Metzger wrote:

> When I asked one of the high cognoscs about the absence of an i/o
> multiplexing primitive in his dialect because I needed one for doing
> some event driven programming, I was regaled with why event driven
> programming is stupid and ugly and I should be using threads
> instead. I've had a dozen such experiences of late.

Isn't this exactly the attitude that Java has (or had, has it changed
now?)

--tim
From: Stephen J. Bevan
Subject: Re: religion
Date: 
Message-ID: <m3bs7c9yc0.fsf@dino.dnsalias.com>
Tim Bradshaw <···@cley.com> writes:
> > When I asked one of the high cognoscs about the absence of an i/o
> > multiplexing primitive in his dialect because I needed one for doing
> > some event driven programming, I was regaled with why event driven
> > programming is stupid and ugly and I should be using threads
> > instead. I've had a dozen such experiences of late.
> 
> Isn't this exactly the attitude that Java has (or had, has it changed
> now?)

It changed in Java 1.4 with the introduction of the nio package.
However, it was added in such a way that it doesn't seamlessly fit
with the other socket abstractions i.e. if you want to run SSL over a
non-blocking socket in Java you are going to have to do a lot of work.
From: Thomas F. Burdick
Subject: Re: religion
Date: 
Message-ID: <xcvbs7cmixt.fsf@hurricane.OCF.Berkeley.EDU>
"Perry E. Metzger" <·····@piermont.com> writes:

> Erik Naggum <····@naggum.no> writes:
> > * Perry E. Metzger
> > | What I've noticed in a lot of people in this community (and it really
> > | is just one community) is a lot of arrogance.
> > 
> >   Funny, that, I see a lot of humility and respect for the opinions and
> >   thoughts of those older and wiser than oneself.  There is virtually no
> >   arrogance among the cognoscenti.
> 
> When I asked one of the high cognoscs about the absence of an i/o
> multiplexing primitive in his dialect because I needed one for doing
> some event driven programming, I was regaled with why event driven
> programming is stupid and ugly and I should be using threads
> instead. I've had a dozen such experiences of late.

I take it you're talking about a Scheme dialect, right?  I would
assume that Erik was talking about the (Common)Lisp cognoscenti, so
that's not particularly relevant.  I have a hard time imagining you'd
get a religious response like this from a Lisp cognoscento, if just
for the fact that this is a programming paradigm question, and someone
who thinks that there is One Correct God-Given Way To Write Code
probably wouldn't be using a mixed-paradigm language like CL in the
first place.

-- 
           /|_     .-----------------------.                        
         ,'  .\  / | No to Imperialist war |                        
     ,--'    _,'   | Wage class war!       |                        
    /       /      `-----------------------'                        
   (   -.  |                               
   |     ) |                               
  (`-.  '--.)                              
   `. )----'                               
From: Michael Sperber [Mr. Preprocessor]
Subject: Re: religion
Date: 
Message-ID: <y9lwuq092qj.fsf@sams.informatik.uni-tuebingen.de>
>>>>> "Perry" == Perry E Metzger <·····@piermont.com> writes:

Perry> Erik Naggum <····@naggum.no> writes:
>> * Perry E. Metzger
>> | What I've noticed in a lot of people in this community (and it really
>> | is just one community) is a lot of arrogance.
>> 
>>   Funny, that, I see a lot of humility and respect for the opinions and
>>   thoughts of those older and wiser than oneself.  There is virtually no
>>   arrogance among the cognoscenti.

Perry> When I asked one of the high cognoscs about the absence of an i/o
Perry> multiplexing primitive in his dialect because I needed one for doing
Perry> some event driven programming, I was regaled with why event driven
Perry> programming is stupid and ugly and I should be using threads
Perry> instead. I've had a dozen such experiences of late.

That would be me (on comp.lang.scheme.scsh), and, until now, I thought
we were having a fruitful exchange on that matter.  I guess I was
mistaken, and I apologize if was "regaling" you in any way.  It
certainly didn't feel that way to me.

You did forget to mention that the same "high cognosc" (I guess that's
a compliment) promised to implement said multiplexing primitive
exactly to accomodate people like you.  In fact, I've actually done
the work by now.

-- 
Cheers =8-} Mike
Friede, V�lkerverst�ndigung und �berhaupt blabla
From: Ray Blaak
Subject: Re: religion
Date: 
Message-ID: <u65xkxv98.fsf@telus.net>
·······@informatik.uni-tuebingen.de (Michael Sperber [Mr. Preprocessor]) writes:
> Perry> When I asked one of the high cognoscs about the absence of an i/o
> Perry> multiplexing primitive in his dialect because I needed one for doing
> Perry> some event driven programming, I was regaled with why event driven
> Perry> programming is stupid and ugly and I should be using threads
> Perry> instead. I've had a dozen such experiences of late.
> 
> That would be me (on comp.lang.scheme.scsh), and, until now, I thought
> we were having a fruitful exchange on that matter.  I guess I was
> mistaken, and I apologize if was "regaling" you in any way.  It
> certainly didn't feel that way to me.

A perfect example of the effective gentle strategy in action: a conflict or
misunderstanding is observed, followed by a response that clarifies or
otherwise smoothes things over.

Of course, the entertainment value is much reduced, but one can't have
everything.

-- 
Cheers,                                        The Rhythm is around me,
                                               The Rhythm has control.
Ray Blaak                                      The Rhythm is inside me,
·····@telus.net                                The Rhythm has my soul.
From: Stephen J. Bevan
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <m31y89gnli.fsf@dino.dnsalias.com>
"Perry E. Metzger" <·····@piermont.com> writes:
> I don't know about that. I'd expect that if the community was truly
> vibrant we'd be seeing things like equivalents to CPAN and such.

By this measure, which languages are truly vibrant?  I'm particularly
interested in any that have multiple (say >= 3) implementations.
From: Perry E. Metzger
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <87admx2ks5.fsf@snark.piermont.com>
·······@dino.dnsalias.com (Stephen J. Bevan) writes:
> "Perry E. Metzger" <·····@piermont.com> writes:
> > I don't know about that. I'd expect that if the community was truly
> > vibrant we'd be seeing things like equivalents to CPAN and such.
> 
> By this measure, which languages are truly vibrant?

Vibrant might be the wrong term. Lets just say "being so actively used
by large communities that lots of tools and libraries are available on
the net".

By that measure, we're talking about C, C++, Java, Perl, Ruby, Tcl/Tk
(Tcl makes my teeth itch but never mind that), Python, and doubtless a
few others I'm forgetting or unaware of.

I'm not passing any moral judgments here, by the way. Just noting
what has become popular enough that you can go out and at will find
yourself books like "Programming the Perl DBI" or libraries to
generate web log templates or what have you. People might discount the
importance of such things, but they're a pretty big part of the way
modern people put together apps. It is a lot easier to steal someone's
RFC822 parser that they've put up on the net than to write one,
regardless of the language, and when there is a critical mass of such
stuff available it makes a difference in your life.

-- 
Perry E. Metzger		·····@piermont.com
--
"Ask not what your country can force other people to do for you..."
From: Kenny Tilton
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3D76D9D7.1060002@nyc.rr.com>
Perry E. Metzger wrote:
> I'm not passing any moral judgments here, by the way. Just noting
> what has become popular enough that you can go out and at will find
> yourself books like ...

Then let's not leave COBOL and Pascal off the "vibrant" list. Remember 
those popular languages?

kenny
From: Perry E. Metzger
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <87fzwp127b.fsf@snark.piermont.com>
Kenny Tilton <·······@nyc.rr.com> writes:
> Perry E. Metzger wrote:
> > I'm not passing any moral judgments here, by the way. Just noting
> > what has become popular enough that you can go out and at will find
> > yourself books like ...
> 
> Then let's not leave COBOL and Pascal off the "vibrant" list.

I'm afraid I must. I can't find large COBOL libraries anywhere on the
net, and there isn't any place to find large Pascal libraries,
either. The same can't be said about C or Python.


-- 
Perry E. Metzger		·····@piermont.com
--
"Ask not what your country can force other people to do for you..."
From: Kenny Tilton
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3D777C0C.1020808@nyc.rr.com>
Perry E. Metzger wrote:
> Kenny Tilton <·······@nyc.rr.com> writes:
> 
>>Perry E. Metzger wrote:
>>
>>>I'm not passing any moral judgments here, by the way. Just noting
>>>what has become popular enough that you can go out and at will find
>>>yourself books like ...
>>
>>Then let's not leave COBOL and Pascal off the "vibrant" list.
> 
> 
> I'm afraid I must. I can't find large COBOL libraries ...


Well, (a) I said "books", but (b) my larger point was that popularity 
comes and goes.

As for your arrogance worries, someone comes along every few months to 
complain about that. It's not arrogance, it is confidence. It just seems 
like arrogance to anyone who does not appreciate how excellent is Lisp.

kenny
From: Stephen J. Bevan
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <m3n0qxezdz.fsf@dino.dnsalias.com>
"Perry E. Metzger" <·····@piermont.com> writes:
> Vibrant might be the wrong term. Lets just say "being so actively used
> by large communities that lots of tools and libraries are available on
> the net".
> 
> By that measure, we're talking about C, C++, Java, Perl, Ruby, Tcl/Tk
> (Tcl makes my teeth itch but never mind that), Python, and doubtless a
> few others I'm forgetting or unaware of.

Of these only C, C++ and Java pass my arbitrary "more than 3
implementations" requirement.  Even there Java stands out from the
other two in that while there are multiple implementations, Sun pretty
much dictates the direction of Java in the same way that the authors
of the single implementation languages do.

That leaves C and C++, neither of which have anything like CPAN. 
Instead you can find various libraries in many places, which vary from
working on almost all implementations to those which work only on
specific platforms or even specific compilers (e.g. Linux kernel isn't
going to compile with anything but gcc anytime soon).

So two possible broad categories of vibrant languages :-

  1. Languages in which a single person/company controls the language.
  2. C/C++

or another possible split is :-

  a. "scripting" language with one (or two) implementation(s)
  b. C/C++/Java

None of the following fit into any category: Ada, APL, Common Lisp,
Forth, Haskell, Scheme, SML and Smalltalk.  All have some archives,
but nothing that approaches CPAN.  There are also some languages that
do fit in category 1 such as Clean, Erlang and O'Caml which also don't
have anything that approaches CPAN (though Caml Hump is pretty
impressive).

The point being that compared to C/C++/Java and a handful of scripting
languages it seems that no languages are vibrant.
From: Marco Antoniotti
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <y6cofbc4kn4.fsf@octagon.mrl.nyu.edu>
·······@dino.dnsalias.com (Stephen J. Bevan) writes:

> "Perry E. Metzger" <·····@piermont.com> writes:
> > I don't know about that. I'd expect that if the community was truly
> > vibrant we'd be seeing things like equivalents to CPAN and such.
> 
> By this measure, which languages are truly vibrant?  I'm particularly
> interested in any that have multiple (say >= 3) implementations.

You realize that you are essentially disqualifying things like Python
and Perl, don't you?

Cheers


-- 
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group        tel. +1 - 212 - 998 3488
715 Broadway 10th Floor                 fax  +1 - 212 - 995 4122
New York, NY 10003, USA                 http://bioinformatics.cat.nyu.edu
                    "Hello New York! We'll do what we can!"
                           Bill Murray in `Ghostbusters'.
From: Stephen J. Bevan
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <m3elc8fot8.fsf@dino.dnsalias.com>
Marco Antoniotti <·······@cs.nyu.edu> writes:
> ·······@dino.dnsalias.com (Stephen J. Bevan) writes:
> > "Perry E. Metzger" <·····@piermont.com> writes:
> > > I don't know about that. I'd expect that if the community was truly
> > > vibrant we'd be seeing things like equivalents to CPAN and such.
> > 
> > By this measure, which languages are truly vibrant?  I'm particularly
> > interested in any that have multiple (say >= 3) implementations.
> 
> You realize that you are essentially disqualifying things like Python
> and Perl, don't you?

Yes, the >=3 is in there deliberately to try and expose any
languages which are vibrant and which also have multiple
implementations.  My theory being that there aren't any.
From: Sander Vesik
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <1031261877.719655@haldjas.folklore.ee>
In comp.lang.scheme Stephen J. Bevan <·······@dino.dnsalias.com> wrote:
> Marco Antoniotti <·······@cs.nyu.edu> writes:
>> ·······@dino.dnsalias.com (Stephen J. Bevan) writes:
>> > "Perry E. Metzger" <·····@piermont.com> writes:
>> > > I don't know about that. I'd expect that if the community was truly
>> > > vibrant we'd be seeing things like equivalents to CPAN and such.
>> > 
>> > By this measure, which languages are truly vibrant?  I'm particularly
>> > interested in any that have multiple (say >= 3) implementations.
>> 
>> You realize that you are essentially disqualifying things like Python
>> and Perl, don't you?
> 
> Yes, the >=3 is in there deliberately to try and expose any
> languages which are vibrant and which also have multiple
> implementations.  My theory being that there aren't any.

Python has:
	* Python itself
	* Jython
	* stackless python

Well, and Jpython but that has been superceeded by Jython

-- 
	Sander

+++ Out of cheese error +++
From: Stephen J. Bevan
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <m3sn0o84bo.fsf@dino.dnsalias.com>
Sander Vesik <······@haldjas.folklore.ee> writes:
> > Yes, the >=3 is in there deliberately to try and expose any
> > languages which are vibrant and which also have multiple
> > implementations.  My theory being that there aren't any.
> 
> Python has:
> 	* Python itself
> 	* Jython
> 	* stackless python

I'm clearly out of date on stackless python since I thought it was
fork of the Python code that re-worked the internals so that it does
not use a stack.  As such it wasn't a from-scratch (re) implementation
and so should (and does) to run all Python code without change.
Python on the other hand cannot run all stackless python code since
the latter has extensions that aren't in Python.

Jython doesn't run 100% of all Python code.  Some differences are due
to incomplete implementation and some due to deliberately different
design decisions (e.g. GC/finalisation) or limitations imposed by
being 100% pure Java.  For example, using chmod from the os module
doesn't work in Jython.  I don't know if and how much these
differences impact the actual portability of Python/Jython code.
From: Christopher Browne
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <al6lac$1nuea3$1@ID-125932.news.dfncis.de>
In an attempt to throw the authorities off his trail, "Perry E. Metzger" <·····@piermont.com> transmitted:
> I don't know about that. I'd expect that if the community was truly
> vibrant we'd be seeing things like equivalents to CPAN and such. It is
> not disgraceful to say "I'm writing an application that needs to
> interface with database X, anyone have a binding written already" or
> what have you. Perhaps the Lisp community is beyond actually having to
> use its language day to day but I doubt that.

I think we _are_ seeing something like CPAN in the form of CCLAN.

The efforts are somewhat less integrated and less coherent than is the
case for Perl (or Python, or Ruby, with their respective communities),
but that isn't _too_ surprising what with there being an actually
diverse set of Common Lisp implementations.  

There is _somewhat_ less demand for CPAN-like stuff for Common Lisp
since the base language has a bit more in it.  

It is _somewhat_ more difficult to construct CPAN-like stuff because
there are a multiplicity of CL implementations.

On the Scheme side of the fence, it's _much_ more difficult to see
common code get deployed for a multiplicity of Scheme implementations
because the base language is so much smaller.  The SRFI process is
nonetheless supplying at least _some_ of this.  And we periodically
hear more about things designed for scsh getting ported more widely
:-).
-- 
(concatenate 'string "cbbrowne" ·@ntlug.org")
http://www.ntlug.org/~cbbrowne/lisp.html
"High-level languages are a pretty good indicator that all else is
seldom equal." - Tim Bradshaw, comp.lang.lisp
From: Bruce Lewis
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <nm98z2gttrq.fsf@magic-pi-ball.mit.edu>
"Perry E. Metzger" <·····@piermont.com> writes:

> What I've noticed in a lot of people in this community (and it really
> is just one community) is a lot of arrogance.

Yes, this is an acknowledged problem.  There's a popular Jesse Bowman
quote, "The problem with Lisp is it makes you so damn smug."

Perhaps you can help.  If there are these other newsgroups where people
are constantly posting with real-world problems that would be hard to
solve in CL or Scheme, it shouldn't be too much trouble to post one or
two such problems here.  Then we would all see how isolated our little
academic exercises are from the real world.

> I've been programming long enough to know that when you're more
> concerned about what the right comment character is than about writing
> good comments you're not on the level of the important any longer.

If you're a language implementor you're talking about something
important.  Most newsgroups don't include discussions among different
implementors of the same language.

> Software, unlike theoretical mathematics, is largely about
> accomplishing things. That leads people to unfortunate mundane
> concerns like finding a module that builds web pages for you or
> finding a module that interfaces to Oracle or finding a module that
> does statistical analysis for you. When a language's devotees no
> longer discuss such pragmatic matters and instead spend all their
> energy on religion, it implies they are not writing code.

It may imply they aren't getting a lot of new blood.  I'm still writing
code, but I already have a module that builds web pages for me and an
interface to my databases.

-- 
<·······@[(if (brl-related? message)    ; Bruce R. Lewis
              "users.sourceforge.net"   ; http://brl.codesimply.net/
              "alum.mit.edu")]>
From: Craig Brozefsky
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <86y9agw5wy.fsf@piracy.red-bean.com>
Bruce Lewis <·······@yahoo.com> writes:

> "Perry E. Metzger" <·····@piermont.com> writes:
> 
> > What I've noticed in a lot of people in this community (and it really
> > is just one community) is a lot of arrogance.
> 
> Yes, this is an acknowledged problem.  There's a popular Jesse Bowman
> quote, "The problem with Lisp is it makes you so damn smug."

Do you perhaps mean Jesse Bouwman aka ·····@onshored.com, and
co-author of IMHO/USQL/ODCL and some other stuff?


-- 
Craig Brozefsky <·····@onshored.com>	             Senior Programmer
onShore Development                       http://www.onshore-devel.com
Free Common Lisp Software      http://alpha.onshored.com/lisp-software
From: Bruce Lewis
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <nm9y9agrvps.fsf@magic-pi-ball.mit.edu>
Craig Brozefsky <·····@red-bean.com> writes:

> Bruce Lewis <·······@yahoo.com> writes:
> 
> > Yes, this is an acknowledged problem.  There's a popular Jesse Bowman
> > quote, "The problem with Lisp is it makes you so damn smug."
> 
> Do you perhaps mean Jesse Bouwman aka ·····@onshored.com, and
> co-author of IMHO/USQL/ODCL and some other stuff?

I don't actually know.  I just remembered seeing the quote a lot, and
googled to find an attribution.
From: Paolo Amoroso
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <wI94Pa7YsPhqZ1nbeVqoEWrahqkb@4ax.com>
[Followup posted to comp.lang.lisp only]

On 04 Sep 2002 16:27:59 -0400, "Perry E. Metzger" <·····@piermont.com>
wrote:

> I don't know about that. I'd expect that if the community was truly
> vibrant we'd be seeing things like equivalents to CPAN and such. It is

Regardless of whether this definition of vibrant is useful or not, there is
something similar ta CPAN for Lisp. It's called cCLan.


Paolo
-- 
EncyCMUCLopedia * Extensive collection of CMU Common Lisp documentation
http://www.paoloamoroso.it/ency/README
From: Will Deakin
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3D75E660.7060502@hotmail.com>
Perry E. Metzger wrote:
> Perhaps we could have a new discussion about why it is that people
> think that religion is more interesting than writing programs.
So then, people or books[1]?

:)w

[1] ...and where do you sit on the call/cc debate?
From: Kenny Tilton
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3D75FEC1.1000301@nyc.rr.com>
Perry E. Metzger wrote:
> I find something interesting about both comp.lang.lisp and
> comp.lang.scheme. In other language newsgroups, like the the ruby
> group or what have you, people spend most of their time asking
> questions about how to solve particular problems they're having with
> their production code using the language or what have you. This tends
> to indicate the people involved are largely interested in programming
> languages as a way of accomplishing their work. In the c.l.l and
> c.l.s, however, we have groups of people who are concerned largely
> about linguistic metaissues...

The same faulty conclusion could be drawn from observing the traffic in 
  comp.instrument.guitar.rock and c.i.g.classical.

:)

kenny
From: Pascal Costanza
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <3D75FCAF.406F8A23@cs.uni-bonn.de>
"Perry E. Metzger" wrote:
> 

> I find something interesting about both comp.lang.lisp and
> comp.lang.scheme. In other language newsgroups, like the the ruby
> group or what have you, people spend most of their time asking
> questions about how to solve particular problems they're having with
> their production code using the language or what have you. This tends
> to indicate the people involved are largely interested in programming
> languages as a way of accomplishing their work. In the c.l.l and
> c.l.s, however, we have groups of people who are concerned largely
> about linguistic metaissues, that is, about their language as a
> religion.

Your comparisons are not fair, for the following reasons.

(1) Programming languages like Ruby are deliberately restricted to a
particular programming paradigm, in that case object-oriented
programming. If you are willing to restrict yourself to a particular
language you have already made several choices, consciously or
subconsciously, and when you discuss things with your peers you won't
get into these kinds of discusssions because you already agree on some
fundamentals.

Lisp is different in this respect because it has a long history of
experimenting with several programming paradigms. Of course you will get
discussions about language features, sometimes heated discussions, if
that's your topic of interest. No surprises there.

(2) Common Lisp and Scheme reveal different design decisions. This is
most likely because their designers had different goals, sometimes
fundamentally different goals. Sometimes people forget that design
issues are never decided upon in a vacuum, but always with certain
concrete and/or abstract goals in mind. More often than not they are
based on hidden assumptions and discussions get heated because the
participators are not aware of their own respective assumptions.

Now assume you had a newsgroup where people of different object-oriented
programming languages would meet. Soon you would get similar heated
discussions. See http://c2.com/cgi/wiki?LanguagePissingMatch for some
examples.

(3) comp.lang.lisp is about several Lisp dialects. Imagine a kind of
comp.lang.commonlisp - I assume that in such a newsgroup the outcome
would be different. What I want to say here is that comp.lang.lisp is
not about _one_ language but about a family of languages. Again, it's
not surprising that you get discussions about the differences between
the several dialects.

These newsgroups just play in a different league.


Pascal

--
Pascal Costanza               University of Bonn
···············@web.de        Institute of Computer Science III
http://www.pascalcostanza.de  R�merstr. 164, D-53117 Bonn (Germany)
From: Thomas Bushnell, BSG
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <87bs75l2ih.fsf@becket.becket.net>
"Perry E. Metzger" <·····@piermont.com> writes:

> Perhaps we could have a new discussion about why it is that people
> think that religion is more interesting than writing programs.

Of course, you mean PL religion, here, not actual religion religion.

People may just think that newsgroups are a better medium for PL
religion than for writing code.  
From: Thomas Bushnell, BSG
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <87fzwhl2kl.fsf@becket.becket.net>
Al Petrofsky <··@petrofsky.org> writes:

> There are certainly some readers of comp.lang.lisp who are interested
> in comparing scheme and cl macros, but those people should all be
> reading comp.lang.scheme as well.  Scheme is not a recent offshoot of
> lisp.  Lisp followers have had plenty of time to consider whether or
> not they are interested in the directions scheme has taken, and those
> who are interested should look in comp.lang.scheme.

Almost right.  

There are certainly some readers of comp.lang.scheme who are
interested in comparing scheme and cl macros, but those people should
all be reading comp.lang.lisp as well.  Common Lisp did not only
recently diverge from Scheme.  Scheme followers have had plenty of
time to consider whether or not they are interested in the directions
Common Lisp has taken, and those who are interested should look in
comp.lang.lisp.

Thomas
From: Biep @ http://www.biep.org/
Subject: Re: What's gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <alt98r$uecc$1@ID-63952.news.dfncis.de>
Thomas Bushnell, BSG wrote:
> Al Petrofsky <··@petrofsky.org> writes:
>
>> There are certainly some readers of comp.lang.lisp who are interested
>> in comparing scheme and cl macros, but those people should all be
>> reading comp.lang.scheme as well.  Scheme is not a recent offshoot of
>> lisp.  Lisp followers have had plenty of time to consider whether or
>> not they are interested in the directions scheme has taken, and those
>> who are interested should look in comp.lang.scheme.
>
> Almost right.
>
> There are certainly some readers of comp.lang.scheme who are
> interested in comparing scheme and cl macros, but those people should
> all be reading comp.lang.lisp as well.  Common Lisp did not only
> recently diverge from Scheme.  Scheme followers have had plenty of
> time to consider whether or not they are interested in the directions
> Common Lisp has taken, and those who are interested should look in
> comp.lang.lisp.
>
> Thomas

There is a general group for the Lisp family, comp.lang.lisp, and then
there are groups for individual members or subfamilies.  In the general
group there is enough CL-specific traffic to warrant a CL-specific group,
but as no such group has been created, it is acceptable to use the general
group for these discussions.

Discussions about Lisp in general, or comparative discussions that do not
fall within the charters of one of the more specific groups, belong in the
general group comp.lang.lisp.

Now it may be that, to respect the sensitivities of the CL community, such
discussions are actually held elsewhere, but if so, that is out of
courtesy, not because that would be according to the rules.

--
Biep
Reply via any name whatsoever at the main domain corresponding to
http://www.biep.org
From: Will Deakin
Subject: Re: What's gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <altdlj$8rn$1@venus.btinternet.com>
Biep @ http://www.biep.org/ wrote:
> Now it may be that, to respect the sensitivities of the CL community, such
> discussions are actually held elsewhere, but if so, that is out of
> courtesy, not because that would be according to the rules.
Hmmm. I was under the misaprehension -- short of c.l.l.[1] becoming 
moderated[2] -- that there were *no* rules governing what is courteous 
or anything else, really.

:)w

[1] and c.l.scheme et cetera, et cetera....
[2] ...the Saints preserve us...
From: Kaz Kylheku
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <cf333042.0209030727.1b3454ff@posting.google.com>
····@swirve.com (Adrian B.) wrote in message news:<····························@posting.google.com>...
> ····@ma.ccom (Takehiko Abe) wrote in message news:<·····················@solg4.keke.org>...
> > In article <····························@posting.google.com>, ····@swirve.com (Adrian B.) wrote:
> > 
> > > Is there something fundamentally broken with Scheme macros? 
> > 
> > Please do not cross post to comp.lang.lisp.
> 
> And why exactly not?  Lispers are strong users of macros and may have
> some good experience on the design and use of macro systems.  Its not
> like the question was crossposted to a C++ or Java newsgroup.

Actually it's exactly as if the question was crossposted to a C++
or Java newsgroup.
From: Nils Goesche
Subject: Re: What's gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <87wuq5hwey.fsf@darkstar.cartan>
····@swirve.com (Adrian B.) writes:

> Is there something fundamentally broken with Scheme macros?

Yes, they're eugenic.  However, I think there is no need to
discuss this in comp.lang.lisp.  Follow-ups set.

Regards,
-- 
Nils Goesche
Ask not for whom the <CONTROL-G> tolls.

PGP key ID #xD26EF2A0
From: Marco Antoniotti
Subject: Re: What&#8217;s gone wrong with Scheme Macros? Why all the debate?
Date: 
Message-ID: <y6c3css1g5h.fsf@octagon.mrl.nyu.edu>
····@swirve.com (Adrian B.) writes:

> As a casual user of scheme and reader of the related newsgroups, I've
> noticed that there's always so much debate, confusion, and critiques
> of the scheme macro system.
> 
> Is there something fundamentally broken with Scheme macros?  Are they
> badly designed?  Or is all this discussion simply due to confusion?
>  
> Why the constant debate?
> 
> If the macro system is badly designed, are there any serious attempts
> to improve it?   Or is it cast in stone forever more due to appearing
> in the R5RS?  I've seen one alternative, syntax-case, but is it a
> serious alternative for the standard, or is it just a slightly better
> version of syntax rules?


There is no real debate.  Scheme needs the macro system it has because
it is useful and because of the conflating of "variable" and
"function" namespaces.  Not having a macro systme would impair Scheme
usefulness.

Another matter is the pointless debate about the "superiority" of
Scheme "pattern directed" macro system over the "structure handling"
macro system of Common Lisp.  Even in this case the chase can be cut
short in many ways.  It is simply a non problem.

Cheers

-- 
Marco Antoniotti ========================================================
NYU Courant Bioinformatics Group        tel. +1 - 212 - 998 3488
715 Broadway 10th Floor                 fax  +1 - 212 - 995 4122
New York, NY 10003, USA                 http://bioinformatics.cat.nyu.edu
                    "Hello New York! We'll do what we can!"
                           Bill Murray in `Ghostbusters'.