On Wednesday, December 14, 2005, at 11:25 pm, John Pallister wrote: > Hi, >
--->8--- snipped --->8--- Hello there... > > FWIW, I would also like to get into McCLIM (and Climacs), rather than, > say, > porting SWT. I had the pleasure of meeting and listening to Robert > Strandh > earlier in the year, and I'm convinced it's the way to go. Or was, > until I read > Duncan Rose's post, and now I'm very keen to learn more about his > "fourth > generation" CLIM. We've got Chris Double and Bruce Hoult in this town; > I could > try and persuade them to bring their combined Lisp & Dylan experience > to bear. ;) > The best place to learn about DUIM is by searching c.l.l for posts matching "SWM CLIM" or "SWM DUIM". There's a bunch of stuff on the fundev and gwydion dylan websites about the Dylan implementation -- including documentation -- a the moment my port parallels the Dylan code. I wouldn't want to stop people looking at McCLIM; a lot of good work has gone on there. DUIM is far (far, far) less functional than CLIM. There are no presentation types (well... that's not quite true...), it's not stream based (i.e. an arbitrary window is *not* a Lisp stream), there's no text input command loop by default (no input editing!) and several CLIM pane types are missing. Compared to CLIM, DUIM is significantly simplified. Some of the primitives are modernised (DUIM supports path-based drawing; CLIM does not for example). One of the design goals of DUIM was to be simple enough for your average developer to hold everything in his or her head. The layered architecture is a reality in DUIM rather than an ideal (personally I think there's quite a lot of cross-layer contamination in the CLIM spec which means much of the code is more tightly coupled than it should be). [to be fair there are places even in DUIM where it's not clear where specific functionality should live so some cross-layer contamination persists still.] In short: CLIM sings, dances and washes your dishes but might take a while to convince to do so. DUIM only waits tables but does so promptly. Besides that McCLIM is available now and DUIM won't be for a while. I've been blogging about my progress over the last couple of months at http://www.robotcat-blog.blogspot.com if you're interested. I'll try to gather a few pertinent links together and post them there tomorrow. -Duncan ps. I think just this GUI discussion is enlightening as to the 'state' of Lisp in many ways. In just a week we have talk of wxWidges, ltk, gtk, McCLIM and DUIM. Nobody has talked about CAPI or the Franz toolkits (yet) and there are probably other projects out there to surface. This does of course lead to a combinatorial explosion of possibilities in terms of support; is it realistic for apps to be both Lisp implementation neutral and windowing toolkit neutral? I suspect it isn't, although this obviously isn't such a problem for libraries as opposed to apps with a GUI (just a linear explosion for those (proportional to the number of implementations). -D > Cheers, > > John :^P > > [1] http://wiki.alu.org/John_Pallister's_road_to_Lisp > [2] http://lemonodor.com/archives/001305.html > -- > John Pallister > [EMAIL PROTECTED] > > > > _______________________________________________ > Gardeners mailing list > [email protected] > http://www.lispniks.com/mailman/listinfo/gardeners > _______________________________________________ Gardeners mailing list [email protected] http://www.lispniks.com/mailman/listinfo/gardeners
