Brian Mastenbrook bmastenb@cs.indiana.edu writes:
I know fetching from CVS is easy, but that's not the point here. The fact that there has been no release yet send the wrong message, I think.
Not that I disagree, but to play devil's advocate development still seems to be changing the code base fairly quickly. While SLIME is certainly useable right now (depending on your definition of useable), the rate of change in the codebase suggests instability.
A minor pet peeve I have about many free software (or open source the non tm version) projects is the 0.x releases that never seem to make it to 1.0 (or just haven't yet). I know that is a feature, not a bug. But I think that the SLIME project is doing something rather courageous by not cutting an early release.
Right now, it feels more like XP. Release early and often. This is the time for user input to iron out the major issues before a real 1.0 release. Of course more people with a sense of adventure to run the CVS version (HEAD, not FAIRLY-STABLE) would be a good thing if you buy the theory that many eye balls make bugs shallow.
That said, SLIME is quite useable. Perhaps it is time to consider what needs to happen before the SLIME maintainers are willing to slap a version 1.0 or at least a version 1.0 RC1 tag onto CVS. Most of the supported Lisps are not fast moving targets. So if the features are there, then its just the nastier bugs that need to be stomped.
While I run Emacs from CVS on Mac OSX, I run the version in Debian/testing on Linux. I would recomend slime.el being functional with the latter version as it is the current "stable" version. The same would apply to XEmacs compatibility. GNU Emacs in CVS is a moving target and it shows. Or at least it introduces even more variables into the equation.