On Thu, 10 Sep 2009, Aaron W. Hsu wrote: > On Thu, 10 Sep 2009 04:28:58 -0400, Ray Dillinger <[email protected]> wrote: > >> Can we at least agree that a developer, having read the Thing One >> report, should be able to *use* the module system without making >> guesses as to how a particular implementation organizes it? R6RS >> failed to address finding standard modules in an installed system. > > It failed to address this for good reason. A lanugage standard is a > standard for a language, an Thing One should be first and formost a > language document, not some document about how to implement this language. > Identifying how to find and load libraries necessarily restricts the > domain of use to systems that *have* the mechanism described. It makes no > sense to take very implementation specific features and stuff them where > they do not belong.
Yes, in this regard I must say I like the R5RS agnosticism about how programs are represented and implemented. The wording of the R5RS programs section still applies even if the program is written on the back of an envelope and executed by interpretative dance. I remember myself and others really fighting to keep posix-specific stuff out of the R6 toplevel syntax, for example. IMHO these things have no place in a standard. Andre _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
