Ali Chamas wrote:

> I had a quick peek, and it sounds cool, how much has
> been done, what does it do?

The Email model building framework is in pretty good shape.
On the data side, I have every message since November 96.
I'm actually working on it again right now, building a visitor
framework to support various tools that would traverse the email
model while gathering various bits of information.  Something
that would be a big help would be to have an IMAP-compatible mail
client that gets the messages off of the mail server.  A POP client
would do, but IMAP would be better in my environment.  I'll
probably retrofit a POP client from one of the existing Smalltalk
email frameworks for this purpose. Once I have the messages, my
framework can then build the models to be used by other tools.

I also have a good start on a framework for parsing Squeak's
change files and file-outs ( you might ask why I'd do that-
simple... Squeak's own framework is broken ).  On the data
side, every source and change file of every release since 1.1 .
I use that framework within Oasis- so far, it handles all
the files that I've been able to feed it.  It generates chnage
objects, and I have a visitor framework started on those as
well.

Another major information resource would be web pages-
of which wikis are a subset.  There are already several
parsers available written in Smalltalk, but html has really
become quite a bear in the last few years.  However, wiki
pages are very simple compared to the general case web page,
so it should be possible to build something to integrate
wiki content as well- but I haven't done anything along those
lines yet.

I haven't worked out the specific details of how I'll link these
information resources together, but I've done similar things
in the past.  Bigger problems are handling the evolution of the
thing- so for instance, the email resource needs to update continuously
whenever possible, while source code has it's own set of problems.

That still leaves an awful lot of work to be done, though.
However, there would be lots of useful intermediate levels
of ability as well- no need to wait for the grand solution
to have lots of benefit.  For instance, one step would be
to integrate the above into a wiki.

The same framework could support any of the other Smalltalk
communities as well- there would be a Squeak-specific solution,
but that solution could then also be used by other dialects.
That's why my own efforts are more strongly oriented towards
emphasizing cross-dialect useability.

> There has been talk for a while now about merging all
> of the support efforts of the Squeak community into
> one central place. I have personally suggested a
> built-in area of Squeak which should support this.

Basically, the information resources should be an extension
of the resources we normally have within the image.  There's
an awful lot of tool and infrastructure development involved
in building something like this- what I've done so far is
barely scratch the surface.

> I think that if we moved all of the Swiki/Wiki misc. web
> sites, documents, the mailing list, and the likes,
> into the actual image of Squeak (which could then be
> purged, the same way that Squeak can shrink other
> things), then people would not have to hunt down
> information, it's right there in their development
> environment.

Yes- I think that is the correct virtuality.  The actuality
I would expect to be different, but to the user that should
not be an issue.  And that is where lots of tough problems lie.
I also have a wiki dedicated to the Xanadu source code- rather
than describe what I think, have a look at what others have
already wrestled with. ( change Squeak to Xanadu in the prior url ).

> Plus, it would also boost Squeaks
> capability basis, and would be a good demonstration of
> the abilities of this Language/Environment (plus all
> the other current "proof" that the world needs).

Smalltalk's been proving itself for over two decades.
I don't see this as an issue- over time, more and more
of the things we take for granted are finally being accepted
by the mainstream.  The bigger issue for me is making the
Smalltalk world more readily accessible to newcomers as well
as myself.  I see this sort of system as being aimed at solving one
aspect of that problem.

Finally, I'll leave you with another thought- consider
the ramifications of ditching the whole e-mail/wiki/www/etc
mode of communication, and start over from scratch.  What
should a real community information system be?

take care!

- les