On 10 Sep 2009, at 5:06 pm, Ray Dillinger wrote: > I don't care how they are laid out in the file system. [snip] > To reiterate, it's not a question of how files are laid out in the > file system; it's a question of exactly how the language allows > programmers to refer to and use standard modules.
There's three issues here. 1a) As Ray points out, we should be able to request optional standard features in a portable way (which strongly implies without caring about filesystem paths) 1b) However, there's an interesting issue of installation; if we have written a portable implementation of a standard or non-standard library, how is it "installed" into an implementation? This is really down to the implementation, but some recommendations on accepted practice for implementations that work in common platforms might not be out of place. Even if just a recommendation that an implementation supply an "install-module" procedure that takes a filename, and attempts to find a module inside it and install it, or something. 2) We should be able to structure out application as a set of modules that comprise it, in effect a main program plus a bunch of third-party modules, that may be a mix of modules specific to that program (but providing valuable internal segmentation) and bundled third-party portable library modules; and we should be able to put these files in a tarball or something and expect them to work on any implementation with a filesystem (and implementations without filesystems should probably have some tooling to get such directory structures into their implementation's database or whatever). This should be a compulsory part of the standard, just a strong suggestion for filesystem-based implementations; in particular, I think that suggesting that the ability to require a module by name, when that module is contained in a file with the name of the module plus ".scm" and found in a directory alongside the program/module importing it, is a good start; support for absolute or relative paths with directory components, however is a more questionable idea. > > Bear > ABS -- Alaric Snell-Pym Work: http://www.snell-systems.co.uk/ Play: http://www.snell-pym.org.uk/alaric/ Blog: http://www.snell-pym.org.uk/archives/author/alaric/ _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
