Modular Squeak / Update Streams (Re: Smalltalk & Squeak featured on
Slashdot)
David Harris
dpharris at telus.net
Sun Apr 22 23:30:15 UTC 2001
I, for one, am tired of your attitude, it remains overly negative and
personal. I would like to know Squeak Central's opinion on modularity,
whether they see a need for more, if so what, and what kind of priority they
have for those changes.
David
"Andrew C. Greenberg" wrote:
> Les writes:
>
> > When SqueakC announced that they were going to tackle this problem, I
> > immediately signed up on the
> > above Swiki pages. I made my plans clear, and provided pointers to the
> > Wiki sites where I have been
> > describing my work in full public view for over 4 years. I have heard
> > NOTHING, not even so much as
> > a how-do-you-do, from ANYONE. This is not what I expected when I went
> > to the trouble to sign up on
> > those pages, and expose my works in progress to public scrutiny where
> > my thoughts and ideas were
> > free for the taking.
>
> As between SqC's and the Oasis' project's productivity, measured in
> terms of published code, I think it is apparent who has contributed more
> to the Squeak community at this point in time. If there is more to
> Oasis than a few precatory remarks about a project that might someday
> be, I was not able to glean it from the website. Meanwhile, SqC hit
> virtually every one of their marks on their modularity projects. It is
> one thing to shout to the world what you are doing, and how your are
> intending to do it -- and another to show that you are doing it by
> sharing your code. If you want a how-do-you-do, post a changeset or two.
>
> With all due respect, Les, the Wiki site describes some wonderful stuff,
> but that's all it does. As in the world of literature, it is far better
> to SHOW than to tell.
>
> In the open source world, code speaks. Even a deep, well-considered
> design document is useful in this regard. There is a reason for a
> "how-do-you-do," because there is something to speak about. The
> "How-do-you-do"s come to those people who give people a reason to
> respond, or something more directly substantive to which to respond.
>
> At any rate, Squeak contributed and implemented its nameset changes, and
> its project/nameset interactions, pretty much as they set forth about a
> year ago. I was actually quite surprised to hear the thundering silence
> of the modularity advocates, either substantively or materially -- I was
> looking forward to something neat happening there. I, too, was silent
> about it, apart from contributing a few bug fixes and conforming
> modifications to the Slang interpreter. But, hey, modularity wasn't
> that huge a priority for me. Likewise, others posted substantive
> changesets contributing to that project, most of which were incorporated
> into the image.
>
> Why is it that the "silent majority" is better represented in the base
> image on modularity issues than their louder Smalltalk-as-Python
> counterparts? Right, the latter contributed less that could be so
> incorporated. Why is Squeak not yet as they would like it to be?
> Because they haven't offered anything to make it so.
>
> SqC delivered what they described and then promised, and all we heard
> from some quarters is that Squeak still isn't modular enough yet, and
> the "only way" to avoid a fork is for SqC to get more serious about it.
> SqC has the gravitas to make sweeping changes not because they control
> the image, but because they have in fact contributed sweeping changes.
> Others have as well, and many of the other contributions have been so
> incorporated.
>
> >> Which Change Sets improving modularity were made available and not
> >> added to
> >> the upgrade stream?
> >
> > The changes are too deep. This is no simple addition to the image.
>
> I have seen wholesale restructuring of the entire system and image
> encapsulated in a few changesets (the SqC Namespace stuff, for
> example). We have seen ENTIRE REVISIONS of the interpreter, garbage
> collection, code generators and other things demonstrated and
> implemented in changesets. I think the question was well put, and was
> not terribly well answered. Reasonable people may speculate why, but
> if there was a more solid answer, the posted remark, mere pabulum in
> fact, was a poor substitute for it.
More information about the Squeak-dev
mailing list
|