Hi.
That's my first participation here, let's hope it will be meaningful and not
too rude…
My first remark is about the meaning of the steering commitee itself: is it
supposed decide instead of me (as a group of "preferred" representents of
the community), or is it supposed to be the entity that will actually listen
to what I say, and care about it?
I have this feeling that R6RS has been decided disregarding the opinion of
"casual" users, or worse, "potential" users, and that many wasted time
fighting on details that would have been implemented in a way or another
(and potentially fixed later) in other communities.
I am now wondering: what do implementers want? Do they want *their* language
for themselves, or do they want it used?
I want to support the latters. I consider that the formers want as little
features in the langage so that they can implemented what they want, how
they want, and not care about the rest of the community (portability and
joint work).
> Interesting! I feel like, if Scheme had a "industrial features" added
> it would be perfect for industrial users. I don't want to use CL.
>
> Do you think that they can be added without ruining the core?
>
> Why not define the core plus the optional industrial add ons? Why
> can't we have our cake and eat it too?
>
That sounds a bit like what Marc Feeley suggested and like what I ask for.
A small kernel, and optional addons that evolve independantly. Yet, addons
should have *perfect* semantic and API match: it the (winning) API states
"to remove an element, do (hash-set! h x #f) ", they one should not give a
different semantic to this and, say, "(hash-delete h x)".
Any implementation that doesn't support properly the specs should be
considered as not supporting it at all.
This should make compatibility issues clearer.
On the other hand, I don't consider the look of the module system a
compatibility issue. It's like porting from ASDF to mudballs or old
defsystem in CL. Bits of tweaking, yes, but still the same language: my code
will run as it.
To recapitulate my wishes:
- A commitee that cares about the users of the language. All of them. And
that makes voting periods last more than a couple of weeks too
- A small language
- A HUGE set of extensions to whatever has not been set in the language (and
that may involve huge changes in the core interpreter too, I do *not* care).
I would see there non blocking IOs, sockets, threads, complex numbers,
SRFI-1, hash-tables, multiline comments, blah blah
- An breeding pool for everything else: module systems, home-made FFT,
"litterate scheme" parsers to write runnable blog posts. A mix of CPAN,
SRFIs and of gambit's dumping ground. Here, compatibility is no more, but
code is given a chance to be the next buzzword or in the next revision on
the official extensions.
You may not share my point of view, but I think that Scheme should stay
small for the purists, while being (portably) "as bloated as {Ruby, Perl,
CommonLisp}" for those who like the language and yet, want to use it for
practical purposes using their favourite implementation.
So, how many potential members of the commitee would consider these requests
as a potential evolution of Scheme?
Adrien
--
Français, English, 日本語, 한국어
_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
