More on Garnet

Daniel Herring was kind enough to send this query, and I thought I'd make an open response, hoping that it will be of general interest:

Hi Robert,

I read through your TC Lispers slides and was surprised to see Garnet as one of the big three (and no LTK!?!). The garnetlisp sourceforge site looks nearly abandoned, and on cliki Stelios Kokkalis refers to a project which never appeared on common-lisp.net.

After a fair deal of searching, I noticed the garnetlisp CVS tree is still active. You might want to release a current "stable" tarball and update other pages; I thought the project had been abandoned years ago.

Let me take these two points in order:

  1. The material discussed at the meeting --- and particularly mine --- made no pretensions of being representative.  Mine in particular was only about the frameworks that I've spent a lot of time with.  I would very much have liked to hear about ltk, personally, since I have lashed together several crude UIs using tcl/tk and expect/tk, and since it looks pretty good, but no one volunteered to talk about it.
  2. Garnet has been dormant.  Talking about it made me go back and have a look again.  I have been trying to get it working again, using a combination of SBCL and Christophe Rhodes's new, improved CLX.  I chose SBCL, because I thought that would be most appealing to the community at large, but this may have been a mistake --- SBCL had not yet forked from CMUCL at the time most of Garnet was written, so there are a lot of places where the various conditional compilation directives don't cover SBCL.  ACL was active then, so it might have been a better choice.  My wife has been out of town, so there's been no one to complain about me banging on Garnet late at night, but it's not ready for a tarball release yet.  I have most of the tour working, but not all yet.  I will post more on the tc-lispers web site as time permits, and progress justifies.