From: John Cowan <[email protected]> Subject: Re: [r6rs-discuss] Proposed features for small Scheme, part 2 bis: I/O Date: Wed, 23 Sep 2009 01:15:43 -0400
> > WG1 doesn't need to *mandate* unicode. But the slight consideration > > to the possible encodings beyond ASCII, which means just not mixiing > > bytes and characters, allows the code to be much more portable. > > I don't think that's too much to ask. > > I agree. I wonder whether you think that bytevectors/blobs should be > in the core, even though they can be (inefficiently) represented as > general vectors, simply to prevent confusion between them and strings. > (Of course, implementations would in either case be free to provide > efficiently packed blobs.) I've lived with R5RS + srfi-4 uniform vector pleasantly, so I'm inclined to say we can leave them to srfi. If an implementation is capable emulating bytevectors by mutable strings (e.g. in ISO8859-only implementation), it would be trivial to implement u8vector using mutable strings as a backing storage. Portable code doesn't need to break abstraction and still runs efficiently. (It occurs to me that it may be worth to keep a list of items we think we can consider in srfis during discussion.) --shiro _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
