Folks -

The test-pilot version is nearing what I would call 2.5 beta status, following Craig Latta's terminology. We just have a few more outside contributions to funnel in and a bit more work to do on a World-within-a -world window for Morphic. I will announce when I think we're at that point, we'll see what bugs surface for a few days, and then we'll wrap up images and VMs for the release of 2.5.

In order to expedite the completion of 2.5, we decided to sidestep a number of larger projects that are of great importance to many of us. With this message, I would like to return to those larger projects, and move forward with my proposal of 7/13...

>My proposal is that we identify a list of such projects (we have a start already) >and then, for each one, identify a group of 3 to 5 competent people to carry it >through to completion. I'm going to suggest that one of these people be from >Squeak Central, even if they don't actually do any coding. This will ensure that >any concerns we have will get addressed, that we can track the progress, and >that whoever is doing the hard work will know there is buy-in at Squeak Central >from the start. As used above, "to completion" means that it's ready to be >issued as an update to the test pilots.

At the end of this message is a preliminary list of these major areas. I would appreciate it if people would respond DIRECTLY TO ME if they wish to help with

specifing goals, implementation, Q/A

on any specific projects. I will do my best to coordinate formation of appropriate working groups and to keep everyone informed as things take shape. I've never been one to tell others what they can and cannot do, but it will surely reduce traffic on this list (and thus save us all time) if those with a casual interest in the topics below would refrain from broadcasting their many useful suggestions at this point. As each of the various project teams congealeds, they will presumably send out a summary of their intentions and that would be the appropriate time for criticism and suggestions.

I have probably omitted one or two project areas that should be included as formal Squeak Projects at this level. If so, let me know.

Here's the list... -------------------

Exceptions This could be one of the first projects to make it into the base image. There are two reasonably complete implementations one of which is already ANSI compliant.

Block closures Several people have made stabs at this. There are several levels to deal with: accepting the syntax, compiling the code, implementing one or two new bytecodes, and integration with JITTER.

Tim's new compiled method format While Tim seems to have done most of the work already, this involves a major image format change, so we might want to bundle in the other cleanups we have been saving for such a time. Therefore this naturally belongs with a move to 3.0.

Real events from the OS The real goals are: don't miss events, have access to useful other events (like window reframe), and still be fairly simple to port to bare chips. Several stabs have been taken at this. Having caucused with the rest of Squeak Central, we would like to be the inplementors on this, but we welcome suggestions and especially tips on specific platforms. We beg your indulgence until we finish the major competing project.

Proper sockets The major problem is that until now there hasn't been a BSD accept()-like call (which made it possible to miss connections). Update 1328 allows for such a call *optionally* in ConnectionQueue so if you have it then everything is fine and if you don't you just go the way it has been before.

There is a much larger issue of a complete rewrite (along the lines of Craig Latta's Correspondents), and also of a general File System rewrite. We probably need to discuss each of these more fully with regards to pros and cons and urgency. As Craig has pointed out, we will have to move the Mac to OpenTransport when support for MacTCP goes away anyway.

Postscript Support Marcel Weiher has had several working versions of PostscriptCanvases, and we owe him (and all of us) the chance to address any design requests, and finally get it into the base image.

NameSpaces and Packages I had planned to do this, but others in Squeakville have more experience than I. My goal here is to do the simplest thing that provides the major benefits, and tie it somehow into both the image segment loading facility and the new isolated projects capability.

Double-Clickable Squeak Applications, Self-installing Squeak Release I have put these together because the skill set required is pretty similar, and because the second will probably be an example of the first.

Documentation While one might not think of this as a project, it really has all the same surrounding issues of deciding how we want to do it, and coordinating it amongst many contributors, and getting it reall

Tutorials Same comment as documentation

Web Site Same comment as documentation