Survey
A survey was sent out to the contact persons of the 12 stakeholder communities in 2005-09-09. Two reminder emails were also sent out.
Of the 12 comunities 6 answered the sent out emails, but of those only 4 actually filled out the survey. To summarize:
Japanese community, Yoshiki Ohshima: Answered the survey.
SqueakBox and WeatherStation, Dan Ingalls: Answered the survey delegated through Craig.
The Berne Guys, Stephane Ducasse: Answered the survey.
Spoon, Craig Latta: Answered the survey.
Squeakland, Scott Wallace: Answered, but not the survey questions.
Seaside, Avi Bryant: Answered and started filling it out, but got stuck.
Scratch, John Maloney: Answered that he had no time.
SmallLand, Diego Gomez-Deck: No answer.
Croquet, Andreas Raab: No answer.
Tweak, Andreas Raab: No answer.
Georgia Tech, Jeff Pierce: No answer.
Tk4, Steve Riggins: No answer.
This is the original email and below you find the questions with the answers intervowen.
Currently the Squeak community has a temporary leadership called the Coordinators which has among other things started a "work model" based on "teams" (see http://swiki.krampe.se/castaways/6).
- Do you have any feedback to give to us regarding how we work?
Stephane: I have problem to see the real influence or coordinator on the actual
work done.
- Do you have any feedback to give to us regarding the team model currently used?
Stephane: I like the idea that people gather together and get something done.
Yoshiki: What we have would be the least bad system. Stronger leader ship would be better, but we don't have that personality in the community.
Craig (Dan): This setup seems to be working smoothly. There's still the problem of things moving slowly due to a collective lack of spare time, but there doesn't seem to be any cure for that. :)
On 2006-02-15 the Coordinators are stepping down. At that time (or earlier) the Squeak community needs to have a new regime in place.
- Do you have any suggestions or specific ideas on how such a regime should look and work?
- Do you have any specific views on how it definitely should NOT look/work?
Stephane: SqueakFoundation with the coordinators hosted there :)
Yoshiki: I think them stepping down make it more chaotic. I don't mind the Coordinators continue to work on.
Craig (Dan): Each decision should have a deadline attached to it, and someone other than than the decision-maker responsible for prodding the decision-maker about it, preferably in public. :)
Squeak 3.8 has just been released. What are your thoughts on this release?
- Will you start using it, and if not, why not?
Stephane: yes
Craig (Dan):I don't plan on using it. It's still too difficult to move things from one release to another. The Spoon work continues in a 3.2 system, and Weather Dimensions uses a 3.6 system (upgraded from 3.4, at least :). To stay current, I need a better packaging and distribution system. Of course, I have a particular interest here; part of Spoon is Naiad, which I believe to be that better system.
- Do you have any feedback regarding the m17n work that is the major part of 3.8?
Stephane: important
Craig (Dan): From what I hear, it seems to be good work, with a few bugs getting worked out.
- Do you have any feedback regarding how this specific release came about?
Stephane: far too messy and blocked everything for 6 months just because they needed a fast track release. Sorry mike but this was bad.
Yoshiki: There are quite a few projects that have been using it already.
Craig (Dan): I appreciate the diligence people had in keeping the quality up. I think it was worth the extra time.
Squeak is today developed in a release cycle with approximately 1-2 releases per year. The development of the base image is done using an update stream of changesets and the image following this stream is called the "Basic" image. Two more images are defined:
Stephane: There will be no update stream anymore (only for minor actions).
- The Full image which is Basic with extra packages loaded on top. This image is prepared when the Basic image is released.
- The Minimal image which is Basic minus the parts of Basic that are defined as packages using the PackageInfo mechanism. This image is not prepared today, but could easily be.
Yoshiki: Minimal is not necessary, in my opinion. It is waste of time.
As we move more and more code into well defined packages, Minimal willget smaller and smaller, while Basic and Full may very well grow - since we may be adding more useful packages to both. Given that the packages added are easily uninstalled this growth is not considered a problem.
- Do you think the current scheme described above is a good scheme?
Yoshiki: Not really. Making it smaller is not the purpose, but a mean to achieve something. At one point, you have to install something to do any reasonable thing in it. Then, the reason make it smaller beyond that point is not to achieve something, but the size itself will become the goal and people will suffer.
Stephane: Yes. Packages.....
- Would a different scheme suit you better?
Yoshiki: It has been the tradition that alpha, beta, and gamma were made in "Basic", and at the last moment, we made up a "Full" with the gamma basic image and the selected (seemingly important) packages on the web. Probably, we can make alpha, beta, and gamma as future "Full" and remove some package at the last minute to make "Basic" release.
Craig (Dan): I think the basic scheme is fine, although I'd like to see updating happen by synchronization, not by a stream of change chunks. Then one would be more likely to be able to accurately update arbitrary old snapshots with work in them, without having to "move".
There are 9 teams active today. Would you like to start a team around some specific task?
Do you have any feedback to give on the current list of teams?
Stephane: we are going in the right direction
Yoshiki: I can't "commit" myself in regular basis in the Squeak mainstream activity, so being a team (team leader) needs a
little... consideration.
Craig (Dan): I've been working on the task of producing a minimal system and packaging mechanism. It's not to the point where adding people would really help, so it's probably not ready to be a "team", but I'm still working on it. It would probably subsume the "modules" team.
Craig (Dan) regarding current list of teams: Looks good.
The work on Squeak 3.9 is beginning to speed up. There is an evolving plan as we move forward at:
http://minnow.cc.gatech.edu/squeak/5645
- What are your thoughts on that plan?
Stephane: good :)
Craig (Dan): It's reasonable, but I sure would like to see the minimalism issue dealt with first, before putting in any other things (e.g., a new compiler).
- What are your thoughts on the ongoing/performed partitioning?
Stephane: Should continue
Yoshiki: I don't like the idea of MorphicSplitters that trys to cut eToys out.
Craig (Dan): Keep it up!
- What are your thoughts about including Traits in Squeak 3.9? (which is currently looking quite plausible)
Stephane: Important else we simply fork and go to the beach instead!
Yoshiki: Traits is good.
Craig (Dan): To the extent that it doesn't interfere with making the system smaller, I'm for it.
- What are your thoughts about the other larger changes like the new Compiler or the inclusion of the RefactoringBrowser engine?
Stephane: Important!
Yoshiki: RB adds more complexity to be in Basic, if you ask me.
Craig (Dan): I think they should wait until after there is a minimal system.
- What other important things would you like to see addressed in Squeak 3.9?
Stephane: More harvesting. Speed up and removing of junk
Craig (Dan): Nothing. Just make it straightforward to partition and isolate things.
- And what would you like to see further down the road (4.0 and beyond)?
Craig (Dan): Minimalism, modules, new compiler, closures.
We have a "burning issues" list at http://swiki.krampe.se/castaways/2.
-Which of the "burning issues" listed do you think are the most important for us to really get a grip on?
Craig (Dan): "Documentation sucks!"
- Do you have more issues to add to that list?
Craig (Dan): "The system is too big!" This goes beyond the cruft. Most of the system is indeed useful, but should be optional.
And of course finally:
-Do you have any other feedback to give us?
Craig (Dan): Thanks for doing this!
