more on finalization issue, and reference implementation Shiro Kawai (25 Aug 2005 02:39 UTC)
Re: more on finalization issue, and reference implementation Michael Sperber (25 Aug 2005 16:59 UTC)
(missing)
(missing)
Re: more on finalization issue, and reference implementation Michael Sperber (13 Sep 2005 12:25 UTC)

Re: more on finalization issue, and reference implementation Michael Sperber 13 Sep 2005 12:25 UTC

Per Bothner <xxxxxx@bothner.com> writes:

> It would be much appreciated if you could split it up.

Could you suggest a division?

I see four parts immediately:

- primitive layer (including the basic condition hierarchy)
- streams layer (including making transcoder codecs from translators)
- ports layer (including making ports from readers and writers)
- stream ports

The tricky bit is what to do about the transcoders, which are shared
between the stream layer and the ports layer.  Choices are:

- separate SRFI (but not useful independently)
- put it in the primitive layer (but we can't do anything with them
  here)
- put them in the streams layer (but that means the ports layer
  depends on the streams layer)
- put them in both the streams layer and the ports layer (but, if
  there's a 1:1 correspondence between modules and SRFIs, this may
  cause pragmatics difficulties)

If we're just talking about different modules, clearly the transcoders
should be separate.

Suggestions, anyone?

--
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla