Re: rfc (define-module ... #:use-modules ...)

From: Keith Wright <kwright@keithdiane.us>
To: ttn@gnuvola.org
Cc: guile-user@gnu.org
Subject: Re: rfc (define-module ... #:use-modules ...)
Date: Sun, 7 Oct 2007 00:05:13 -0400	[thread overview]
Message-ID: <200710070405.l9745DBe003737@fcs13.keithdiane.us> (raw)
In-Reply-To: <877im0lg1x.fsf@ambire.localdomain> (message from Thien-Thi Nguyen on Sat, 06 Oct 2007 11:59:22 +0200)

> From: Thien-Thi Nguyen <ttn@gnuvola.org>
> 
> for me, it means everyone interested should say what they would
> (or would not) do and then if there is concensus (after some
> refinement), i follow.  if there is no concensus, i muddle
> through the best i can (as always).  from the sound of the
> responses thus far, this is the most likely outcome.

Consensus - same sense or feeling (<-Latin sentire)
Concensus - if that were a word,
            it might mean same head-count

Anyway carry on.  Or muddle on.  My opinion means
little with no code to back it up, and I have none.

>    If you want to harmonize, maybe both branches could
>    think about implementing R6RS library forms.
> 
> the library body is specified to be included in the `library'
> form.  OTOH, `define-module' is a peer top-level form to the
> library body.  how would you reconcile these approaches?

I don't want to rudely inject my opinion, but if
you keep asking quesions it would be rude not to
answer.

I am not totally sure I understand the question.

Are you worried about the systactic difference
between

  (define-module blah blah)
  (def xx)
  (xx xx xx)
  <end-of-file>

versus

  (library blah blah
    (def xx)
    (xx xx xx) )

?

The later (with parentheses on both sides)
seems more lispy to me, but it seems like a
pretty trivial change of syntax.  It would
be more interesting to learn about the deep
magick hidden in the blah blah.

I would leave modules alone for backward
compatibility, and try to add something
with the (library ...) syntax but with an
underlying semantics as much as possible
like the current module system.

Then I would write a paper or manifesto
on the exact reasons why libraries and
modules are too different to be inter-
changeable.

At least, that is how I would reconcile the
two approaches if I were even to begin
reconciliation.  In the real world, I
will type my pipe dream to the mailling
list and then go to bed and not care about
it in the morning.

      -- Keith

_______________________________________________
Guile-user mailing list
Guile-user@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-user