From: Fernando Mato Mira
Subject: Re: What does #+<anything-here> mean?
Date: 
Message-ID: <379D7FD1.602CE873@iname.com>
"Michael Sperber [Mr. Preprocessor]" wrote:

> >>>>> "Fernando" == Fernando Mato Mira <········@iname.com> writes:
>
> Fernando> Barry Margolin wrote:
>
> >> In article <············@nnrp1.deja.com>,  <······@my-deja.com> wrote:
> >> >In article <·················@iname.com>,
> >> >  Fernando Mato Mira <········@iname.com> wrote:
> >> >> Rob Warnock wrote:
> >> >>
> >> >> > By the way, very few Schemes implement #+/#-
> >> >>
> >> >> And it's a very bad situation.
> >> >
> >> >I'd like to know why this capability is considered to be this important.
> >> >(Not meant as a flame.  My only only Common Lisp programming has
> >> >been out of books.)
> >>
> >> It's intended to allow you to write programs or data files that will work
> >> in a variety of Lisp dialects.  E.g. you can write something like:
>
> Fernando> It's not useful only for that. Mainly it's used for differences in
> Fernando> _implementations_.
>
> There are other solutions to that.  You might want to look at SRFI 0
> and SRFI 7 at http://srfi.schemers.org

Some points:

1. This configuration language is a separate stage. That seems to prevent
    adding features dynamically (there's nothing in the SRFI regarding that).
2. Can this be implemented in CL? I'm not sure (maybe by doing something
    horrible like adding reader macros on #\r #\f and #\c).
3. The cascading nature of feature-cond is useful. How could have something
    like that the # way?
4. Add dispatching readmacros to Scheme, and we can have both (I would
    encourage people not to use SRFI7 then if there's no good solution for 2.).

From: Michael Sperber [Mr. Preprocessor]
Subject: Re: What does #+<anything-here> mean?
Date: 
Message-ID: <y9lr9lu48ip.fsf@brabantio.informatik.uni-tuebingen.de>
>>>>> "Fernando" == Fernando Mato Mira <········@iname.com> writes:

Fernando> "Michael Sperber [Mr. Preprocessor]" wrote:

>> >>>>> "Fernando" == Fernando Mato Mira <········@iname.com> writes:
>> 
Fernando> Barry Margolin wrote:
>> 
>> >> In article <············@nnrp1.deja.com>,  <······@my-deja.com> wrote:
>> >> >In article <·················@iname.com>,
>> >> >  Fernando Mato Mira <········@iname.com> wrote:
>> >> >> Rob Warnock wrote:
>> >> >>
>> >> >> > By the way, very few Schemes implement #+/#-
>> >> >>
>> >> >> And it's a very bad situation.
>> >> >
>> >> >I'd like to know why this capability is considered to be this important.
>> >> >(Not meant as a flame.  My only only Common Lisp programming has
>> >> >been out of books.)
>> >>
>> >> It's intended to allow you to write programs or data files that will work
>> >> in a variety of Lisp dialects.  E.g. you can write something like:
>> 
Fernando> It's not useful only for that. Mainly it's used for differences in
Fernando> _implementations_.
>> 
>> There are other solutions to that.  You might want to look at SRFI 0
>> and SRFI 7 at http://srfi.schemers.org

Fernando> Some points:

Fernando> 1. This configuration language is a separate stage. That seems to prevent
Fernando>     adding features dynamically (there's nothing in the SRFI regarding that).

Either a Scheme system implements a SRFI or it does not.  
Dynamicism gives you marginal benefit for maximal cost.

Fernando> 4. Add dispatching readmacros to Scheme,

Read macros are a neat hack, but they introduce all sorts of
conceptual problems.  The experience with having read macros in XEmacs 
has shown that people usually use them in contexts where they are

a) unnecessary
b) cause subtle problems later

I won't even begin to mention the problems that arise when you try to
talk about semantics in the presence of read macros.  It's hard enough 
with the regular R5RS macros.

If you think they're useful, why don't you submit a SRFI?

-- 
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla
From: Fernando Mato Mira
Subject: Re: What does #+<anything-here> mean?
Date: 
Message-ID: <379DA846.501C9F1@iname.com>
"Michael Sperber [Mr. Preprocessor]" wrote:

> If you think they're useful, why don't you submit a SRFI?

http://www.harlequin.com/education/books/HyperSpec/Body/chap-23.html

Done.
From: ····@pobox.com
Subject: Re: What does #+<anything-here> mean?
Date: 
Message-ID: <7nl86l$of2$1@nnrp1.deja.com>
In article <················@iname.com>,
  Fernando Mato Mira <········@iname.com> wrote:
>
> http://www.harlequin.com/education/books/HyperSpec/Body/chap-23.html
>
> Done.
>

An article "SRFI-0 and #+/#- 'clones' as read-time application" posted
on comp.lang.scheme on Jul 16 showed how one can emulate #+ and #- with
cond-expand of SRFI-0 implemented as a read-time application. In fact,
the validation code
	http://pobox.com/~oleg/ftp/Scheme/vcond-expand.scm
makes it a point to use examples of CL's Feature expressions from http:/
/www.harlequin.com/education/books/HyperSpec/Body/sec_24-1-2-1-1.html

The read-time application does many things #-dispatch does, yet in a
somewhat more user-controlled way (stemming from general differences
between application and evaluation).

The article is also available at
	http://pobox.com/~oleg/ftp/Scheme/cond-expand.txt


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
From: Marco Antoniotti
Subject: Re: What does #+<anything-here> mean?
Date: 
Message-ID: <lwoggx9mwt.fsf@copernico.parades.rm.cnr.it>
····@pobox.com writes:

> In article <················@iname.com>,
>   Fernando Mato Mira <········@iname.com> wrote:
> >
> > http://www.harlequin.com/education/books/HyperSpec/Body/chap-23.html
> >
> > Done.
> >
> 
> An article "SRFI-0 and #+/#- 'clones' as read-time application" posted
> on comp.lang.scheme on Jul 16 showed how one can emulate #+ and #- with
> cond-expand of SRFI-0 implemented as a read-time application. In fact,
> the validation code
> 	http://pobox.com/~oleg/ftp/Scheme/vcond-expand.scm
> makes it a point to use examples of CL's Feature expressions from http:/
> /www.harlequin.com/education/books/HyperSpec/Body/sec_24-1-2-1-1.html
> 
> The read-time application does many things #-dispatch does, yet in a
> somewhat more user-controlled way (stemming from general differences
> between application and evaluation).
> 
> The article is also available at
> 	http://pobox.com/~oleg/ftp/Scheme/cond-expand.txt

I should start a web page titled: "Things that Scheme has been
re-inventing after 1984" :)

Cheers


-- 
Marco Antoniotti ===========================================
PARADES, Via San Pantaleo 66, I-00186 Rome, ITALY
tel. +39 - 06 68 10 03 17, fax. +39 - 06 68 80 79 26
http://www.parades.rm.cnr.it/~marcoxa
From: Fernando Mato Mira
Subject: Re: What does #+<anything-here> mean?
Date: 
Message-ID: <379ED2B3.F04322DC@iname.com>
Marco Antoniotti wrote:

> I should start a web page titled: "Things that Scheme has been
> re-inventing after 1984" :)

I knew it! I almost quoted the FTPL myself. Flame on you, now ;-)
From: Fernando Mato Mira
Subject: Re: What does #+<anything-here> mean?
Date: 
Message-ID: <379DA8EC.9408D114@iname.com>
Fernando Mato Mira wrote:

> http://www.harlequin.com/education/books/HyperSpec/Body/chap-23.html

OK, OK. Globals need to be replaced. We need MzScheme parameters
standardized first?