Hi Goran -

> A good article :) on it written many, many eons ago:
>
> http://anakin.bluefish.se:8000/sqworld

Talk about a blast from the past! This is still a pretty good article btw. 
And a few of those questions are just hillarious. Say "(if Stable Squeak 
shold become more "open":) What do you propose as the first steps in opening 
up Stable Squeak more?" Hey, wasn't I just hearing this from you? Maybe 
there's a pattern here ;-) (though I would hope that Tweak doesn't end up 
with the same fate) Or this one "Getting this beast out the door." - seems 
to be the motto of the day :-) And this is a good one, too "What Joseph 
showed me felt like "the future" for Squeak", well no sir, you got that one 
wrong :-) (seriously, this is a very funny read in 20/20 hindsight)

Well anyway, thanks for sending the link - now if I could just dig up the 
actual source code to have something to start from...

> The tricky part with all this might be to find someone with the "itch".
> I mean, most frameworks have someone holding the torch, like you for
> Tweak or Ned (kinda) for Morphic. But who would hold the torch for the
> abstract layer on top?

Anyone who a) wants his or her framework to be Very Easily Usable (tm) with 
anyone's tools (assuming the tool maker cares about it) and b) isn't 
alltogether happy with having to re-hack about five dozen apps in the 
(known) Morphic universe (browse the extensions of UIManager to get a feel 
about the ugly places that you have to poke around in).

In other words, anyone who has had the need to even provide rudimentary 
replacement support for PopUpMenu and friends - I would suspect that 
includes at least wxSqueak (and perhaps Seaside, depending how serious Avi 
takes remote debugging). That's why I said to get these people into the same 
room and see what they need - if there's a need, there will be action. I 
know that (personally speaking) I would go with *any* approach which gives 
me the tools back and I'm almost sure the wxSqueak guys would love it if a 
new tool could just come up with a native UI.

> As most often is the case not much will be done if noone has that itch.

Heh. Exactly. Read my message as a search for collaborators. Together we may 
even have a chance to master the process of getting approved into the image.

> Btw, if we are going for an XML description (or hey, why limit us to ONE
> way of expressing UIs) we could simply try to use the same that say
> Glade uses. Or the one wx uses. Or at least look at them so that we can
> interoperate.

I don't care. I really don't. All I need is TSTTCPW. If that's XML (because 
we can reuse stuff from SWT) so be it. If that's sort of an AppRegistry with 
a well-defined interface so be it. If that's Glade, Array specs or anything 
else that someone else is happy to do ('cause I'm only do one of the former 
two if I'm going to be the lead) so be it. I just need TSTTCPW.

That's another major point when you really have an itch to scratch - you 
stick with the needs. For example, could I have provided more elaborate 
protocols with UIManager? Sure I could - but I don't *need* it. What is in 
there is what I really *need*. No more, no less. TSTTCPW. And I would like 
it to be that way with cross-framework UI support as well. KISS. TSTTCPW. Or 
whatever else you call it ;-)

Cheers,
  - Andreas