On Thu, Feb 19, 2009 at 02:14:12PM -1000, Shiro Kawai wrote: > But in general, I agree that providing binding to existing GUI > widget libraries is far more practical than trying to specify > some 'neutral' API. Practical GUI library tend to require large > set of APIs, and it keeps changing (I've started from Xaw and > OpenView, then Motif, then Gtk1&2, along Java awt and swing). > If we try to have somewhat more 'abstract' APIs separate from > existing toolkits, it would be either too simple common > denominator or too isolated tower of API (I imagine CLIM; it is > an impressive work, but few seem to be using it).
Speaking of which, has anyone had any success with something like gtkserver, or the equivalent for swing (which I don't know for sure exists, but would guess that it must)? Is the notion of running your widget set in a separate process with a bi-directional command stream totally wrong? I don't do GUIs myself, yet, but have done some graphics on a Tektronix terminal, long ago. Perhaps, to be useful, this sort of approach amounts to an API wrapper library anyway, but backed by a less efficient mechanism than the CFFI direct-call approach? Cheers, Andrew. _______________________________________________ r6rs-discuss mailing list [email protected] http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
