Squeak VM, real-time I/O, and threading
Craig.Latta at NetJam.ORG
Wed Jan 30 16:07:00 UTC 2002
(Phooey... I sent this message prematurely before...)
('Sorry for the narrow quote width... I'm on a handheld and haven't implemented automatic quote-filling yet:)
> What about license then? Heard
> some rumour before that Flow is
> under LGPL?
(John Sarkela had asked me last year how I wanted to licence Flow, and I answered LGPL.)
> I downloaded the latest zip but
> found no license at a cursory look.
'Sorry about that... I didn't specify licensing details because I hadn't decided (and I was comfortable with the default legal standing in the interim).
> If I remember the latest "license
> threads" correctly we (the
> community) pretty much agreed
> that stuff that counts as "base
> image" (when modules arrive; I
> guess some modules will be
> officially dubbed as "base"
> modules) should be released under
> SqueakL (at least).
Currently, I lean toward licensing Flow under the recipient's choice of the Squeak license, LGPL, or a license I plan to write that more accurately expresses my desires (I have no qualms about adding yet another license to the universe :). So I assume Flow will enter Squeak under the Squeak license.
> I like Flow a lot even if I have a few
> itches about it (only nitpicks but
> anyway, as the author you
> probably want any feedback you
> can get):
> 1. I personally don't like the added
> methods in Boolean like
> #yourselfUnlessFalseDoFirst: . I
> had to stare and check and think
> and almost still didn't get it! ;-) My
> advice: Get rid of those. :-)
Hmm, they seemed straightforward and useful to me at the time. :) I'll take another look...
> The Client/Server classes built for
> reuse by inheritance would be
> easier to use, I think, if
> constructed for reuse by
I think that's probably true... I tend to defer such specialization until the complexity gets to a certain point, and I felt I hadn't reached that point with the client/server modeling yet.
> And just to make things clear, the
> reasons I abandoned Flow was that
> it ISN'T in the base image and thus
> currently suffers from that.
Right... Although I'm interested in making Flow convenient to use for everyone, I'm encouraged by the potential for a distributed module system to make inclusion in a particular snapshot less of an issue. I think modularity will also help the sync problem too.
More information about the Squeak-dev