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] |
>> What was the reason not to use mzscheme, which has had all >> these features and a "full numerical tower"? Jim> Didn't mean to avoid your question. I don't know. I don't Jim> know much about mzscheme. Tom Lord originally proposed SCM to Richard Stallman, and Richard's advice came from a small group of people. Some of the advice (about language translators) was outright BS, claiming that a Tcl->Scheme translator had already been written, which is not true at all. The MzScheme and RScheme groups proposed their interpreters as GNU project extension interpreters, but Tom and Richard blew them off. It can be argued that these other Scheme systems were more adapted to Guile's goals (they had threads and a much more orthogonal C/C++ interface from the start, and already had many of the "industrial strength" features that many real Scheme practitioners needed). One of SCM's pluses (*extreme* portability) is lost in Guile, and there are no plans to recover it. Another (fast speed for non-native-code systems) is nice, but performance has never been my real preference. But now Guile has picked up a lot of features, and will pick up more, and will go through a clean-up phase. It is definitely getting to be the kind of system you would choose to do your work. I think there is a lesson to be learned (the FSF way of making choices -- sometimes trusting real losers and also pissing off good solid hackers -- is frequently stupid) but not much else to be done. One thing that can be done is to promote the portable C interface, but I need to do more work on that before I can urge the RSchem, MzScheme and STk implementors to adopt it: the gh_ layer is in a bit of a state of disarray right now.