This is the mail archive of the guile@cygnus.com mailing list for the guile project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Date: Fri, 30 Jan 1998 10:52:19 -0500 From: Jim Blandy <jimb@red-bean.com> Cc: brown@bibliotech.com, guile@cygnus.com Sender: owner-guile@cygnus.com Precedence: bulk Content-Type: text/plain; charset=us-ascii Content-Length: 1045 I sympathize with the desire for a consistent syntax. I'd suggest writing something afresh that parses a subset of Guile's read syntax. >i think the point is that we're already dealing w/ a library. the >question is how efficient is its partitioning. (i must admit, it's not >a question that i think deserves too much manual study; i guess i would >appreciate pointers to source-code analysis systems, preferably free.) It seems to be traditional (some excuse!) for lisp systems to be incestuous. i.e. you ask for one little exception-handling function, which brings in the formatter, which brings in the complex number support, and the I/O port implementations, ... Scheme 48 seems to have paid a lot of attention to clean partitioning (e.g. the papers claim you can bring in only the numeric types you need, and everything will just work), but Guile hasn't. You'd have to separate read.c from the allocator. You'd have to separate read.c from all the datatypes you don't care about reading (uniform vectors, multibyte strings, etc). Why can't people just use guile? If the application is large, it probably should have guile embedded in it. If the application needs a configuration file, but should remain small, it can fork guile with the configuration script and another small script which would write the results to a pipe. If the application is really small and light-weight, it probably should just be a guile script anyway. Guile isn't that large, and ideally, the core library should be loaded only once, anyway. If loading a wee bit of guile starts loading huge amounts, then the guile module structure should be fixed. -- Clifford Beshers Computer Graphics and User Interfaces Lab beshers@cs.columbia.edu Department of Computer Science http://www.cs.columbia.edu/~beshers Columbia University