SqC, SqF, and Sq itself
Dan Ingalls
Dan at SqueakLand.org
Fri Aug 10 03:35:32 UTC 2001
--
--
Folks -
A number of people have asked outright what's going on with Squeak
Central and others may be wondering the same but not asking.
I thought I would take a little time out to bring everyone up to
date, not only about to SqC itself, but also on some topics that
spread out into larger community goals.
As you know, Squeak Central has reconstituted itself as an
independent research group. We are talking with a number of possible
funding sources as well as with various potential collaborators, but
it would not be appropriate to talk about details at this time. The
important thing is that we (and it is mainly Alan who is carrying
this process) do not want to rush into any imperfect relationships,
so it is taking some time to set up a situation that works for
everyone involved. I can tell you that our priorities remain...
staying together,
making a real contribution to science and education,
continuing to work with the Squeak community,
sharing our work actively and openly.
Of course it is often a bit harder to focus during times of change,
but on the technical side, we've been putting energy into...
sundry details of enabling Squeak content to load
successfully over the web into any browser
running on any OS on any computer,
continued work on the scripting system and its
relationship to squeak as a whole so that
someday it can all be an organic whole, and
yet it won't seem daunting at first encounter,
increased support for composition of media within
Squeak, seeking that "sweet spot" where things
are still simple, but the capability spans
most day-to-day needs while including the
leverage of an integrated programming environment.
You'll see some of this work in the next spate of updates (today or
tomorrow), and probably more over the next month.
After issuing the pending SqC updates in the next day or two, I will
contact the "iron men" who have helped in the past to see if they
would be willing to take a pass over recently submitted updates,
selecting those that seem solid, valuable, and making sure they work
with the latest updates. [This has been Stefan and Henrik in the
past, as well as Bob Arning and others in SqC. If you think you
would be good at this, and would like to help out, by all means send
me a message offline]. Finally, when they are assembled, I'll issue
these other updates.
The process of judgement, QA, and integration of fixes and
enhancements is crucial to the ecology of an evolving open system.
It is a systemic contribution whose value is probably greater than
any single clever algorithm in the system. While this has been
handled informally in the past, either through my sporadic bursts of
responsibility, or with the help of other committed souls, I feel
that this would be an excellent early focus for SqF. It takes work,
but not too much, it needs to weigh the greatest good for the
greatest number in the system and in the community, it advances the
system, and it enhances the experience of a system that grows and
improves with each contribution from its community.
The importance of merging contributions from the community brings to
mind the need for a package architecture in Squeak. Many people have
already pointed out this need (it's even a current topic), so I don't
think I need to argue the case. The benefits would include:
a framework for ensuring the composability of the
various subsystems that make up a complete squeak image.
the ability to pick and choose among various facilities
and configure an image with these and little more.
the ability to build minimal systems that are well-behaved
with regard to facilities that are missing
But perhaps the most important benefit for the community would be to
reduce the "tyranny of the monolithic image" by which too little gets
included in the release for fear it would be too hard to strip out
(Connectors, Cassowary, MathMorphs, for example), and too much gets
included in the release for fear that if it is not included it will
get orphaned as the system evolves forward (VM Construction,
Wonderland, Speech Synthesis for example).
I propose that we start a thread on this topic (message already sent,
though it was intended to follow) to assess what we have to start
with, and who might be willing to carry something through to
completion. As with many facilities in Squeak, I'm hoping we can
strike a reasonable compromise whereby most code looks no more
complicated, the kernel as delivered is well factored, people who
just want to hack will not be burdened, but those who wish to produce
well-behaved packages will have simple but adequate tools at their
disposal.
Hopefully, just by coming together around this particular need, we'll
generate some good cooperative energy. But I think this project is
also special in that, when completed, it will be the context for an
entire future of better sharing and collaboration around Squeak.
That's it for now
- Dan, for all of us at SqC
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20010809/00df34bb/attachment.htm
More information about the Squeak-dev
mailing list
|