[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: more on finalization issue, and reference implementation

This page is part of the web mail archives of SRFI 68 from before July 7th, 2015. The new archives for SRFI 68 contain all messages, not just those from before July 7th, 2015.

Michael Sperber wrote:
Sure.  The specification is carefully organized in sections so as to
make such a split possible.  In fact, the reference implementation is
organized as layered modules in accordance with the division in the
draft.  The main reason I didn't write three SRFIs is that I wanted to
develop them in close tandem, not because I disagree with the
organizational principle.

It would be much appreciated if you could split it up.  One reason is
it allows implementations to more easily declare they implement
(say) the top-level, without confusion.

I also think a stand-alone ports layer would be easier to add to
R6RS as a modest extension to R5RS.  The streams layer is a separate
library.  The primitive i/o layer is useful (one reason, as
you say, is it provides a hook to define new types of ports), but
it probably should be an optional module/feature that small Scheme
implementations might leave out.

Finally, separate SRFIs would make discussion easier and clearer.
If it is three layered modules that can be independently implemented,
it should have three separate specification documents and three
separate mailing lists.
	--Per Bothner
per@xxxxxxxxxxx   http://per.bothner.com/