| Donal K. Fellows 2007-02-21, 8:05 am |
| Patrick Finnegan wrote (on behalf of Charles Oliver Nutter):
> (Hi there Tcl enthusiasts and Jacl devs! This is kind of a form letter,
> but I'm serious about getting this event together with devs from
> multiple languages. I'd love to hear your thoughts on such an event and
> see if you'd be interested in attending.)
I'd like to really encourage someone to participate in this - the non-
mainline forms of Tcl implementation are always interesting - but I
cannot do so myself due to JavaOne being far too close to other events
that I'm committed to, alas. (Perhaps someone can go and then provide
F2F feedback to the rest of us in a BoF at the Tcl conference? That'd
be really  , and I'm sure ethanol can be arranged to help
then. :-))
But scheduling issues aside, there's some interesting issues here
anyway. One of the key ones is that many dynamic languages (Tcl
definitely included) use a value model that has _objects_ capable of
changing their type/class dynamically (requiring that the classes that
are switched between satisfy some sort of advanced constraint is
reasonable). I think actions to support this sort of thing would be
valuable.
In terms of GUI stuff, I'd say that Tk's most valuable reusable things
are the geometry manager algorithms. (There are other great bits of Tk
too, but the other parts are harder to work with without Tk itself
there; the GMs are usable on their own.) As such, it might be a
reasonable idea for someone to port the 8.5 geometry managers and get
them into the core Java set; in particular, [grid] is massively beyond
anything in Swing, the AWT or even the SWT. (This isn't JVM work, but
it is something worth reusing in many other places.)
Donal Fellows (with TCT/Tcl Core Maintainer/Tcl2007PC hats on).
|