Re: [Scheme-reports] Draft 3 Comments: Chapter 4 Denis Washington (27 Jul 2011 15:39 UTC)
Re: [Scheme-reports] Draft 3 Comments: Chapter 4 John Cowan (27 Jul 2011 16:57 UTC)
Re: [Scheme-reports] Draft 3 Comments: Chapter 4 Denis Washington (27 Jul 2011 19:00 UTC)
Re: [Scheme-reports] Draft 3 Comments: Chapter 4 John Cowan (27 Jul 2011 19:38 UTC)
Re: [Scheme-reports] Draft 3 Comments: Chapter 4 Andre van Tonder (27 Jul 2011 17:22 UTC)
Re: [Scheme-reports] Draft 3 Comments: Chapter 4 Denis Washington (27 Jul 2011 18:36 UTC)

Re: [Scheme-reports] Draft 3 Comments: Chapter 4 Denis Washington 27 Jul 2011 18:59 UTC

Am 27.07.2011 18:54, schrieb John Cowan:
> Keyword arguments can be done in R5RS without extra machinery, provided you
> are willing to export them as identifiers:  see KeywordsArcfide.

This is a nice trick indeed, but does suffer from some subtle
complications. Consider the usage example on KeywordArgumentsArcfide:

   (define/optional (blah a b c (x: 0) (y: 7))
     (list a b c x: y:))

What if you don't want to call "list", but another procedure which
itself accepts "x:" and "y:" keyword arguments? As both keywords and
symbols share the same namespace, there is no way to refer to "x: the
keyword" separately from "x: the local parameter variable" other than
renaming the former or creating an alias for the latter, which makes the
approach pretty unpractical. A separate keyword namespace wouldn't have
these problems.

However this is clearly a WG2 topic, so it probably doesn't make much
sense to discuss that now.

Regards,
Denis

_______________________________________________
Scheme-reports mailing list
Scheme-reports@scheme-reports.org
http://lists.scheme-reports.org/cgi-bin/mailman/listinfo/scheme-reports