Tim Rowledge tim@sumeru.stanford.edu wrote:
Doug Way dway@riskmetrics.com wrote:
Although it's getting a bit late for 3.6alpha if we're moving to beta on Friday, but who knows...
I _really_ don't think that is a good idea. We really ought to get at least the basic, decently reviewed and tested, stuff that we declared in the image before changing to beta. We're not being pushed by any dumbass marketing droid for a release date so let's take the luxury of at least trying to achieve feature completion!
Sorry, I'm lost. Are you saying that we should wait until someone reviews every single thing in the game plan? Does this include SM1.1 which isn't released yet? and Anthony's run time fixes, which nobody that understands exceptions is volunteering to review?
I think we should release this weekend what we already have finished and reviewed, and see if someone can help along what didn't get ready, so they get in early in 3.7.
If someone can point me to where the code is I'll check out the simulator changes so we can at least get _them_ done.
Cool. Install BFAV from SM, filter by string "simulator".
In return I want someone to review the DeclarativePools stuff. :-)
Too late, AviB has already posted a quite positive review :-) I'll probably look and approve them today or tommorrow.
Daniel
Hi all!
Daniel Vainsencher danielv@netvision.net.il wrote:
Tim Rowledge tim@sumeru.stanford.edu wrote:
Doug Way dway@riskmetrics.com wrote:
Although it's getting a bit late for 3.6alpha if we're moving to beta on Friday, but who knows...
I _really_ don't think that is a good idea. We really ought to get at least the basic, decently reviewed and tested, stuff that we declared in the image before changing to beta. We're not being pushed by any dumbass marketing droid for a release date so let's take the luxury of at least trying to achieve feature completion!
Sorry, I'm lost. Are you saying that we should wait until someone reviews every single thing in the game plan? Does this include SM1.1 which isn't released yet? and Anthony's run time fixes, which nobody that understands exceptions is volunteering to review?
First of all, AFAICT these are the things done from the original 3.6 game plan:
1 Removals 2 KCP 3 MCP 6 mir Network rewrite 7 TrueTypeTextStyle 11 Inclusion of SM plus related packages in the release image (though maintained as packages, not directly by update stream).
Of course, I can't say if KCP/MCP are "done" but at least a bunch of stuff has gone in from those projects. SM is included now - but only version 1.06.
And these are still hanging:
4 Anthony runtime enhancements (split in two - fixes and closures) 5 Craig's simulator fixes 8 Diego look style enhancements 9 Replace fonts with AccuFonts (mainly in order to remove the old - people can now load additional nice fonts themselves anyway). 10 SM 1.1
Out of these I think all are "built" except for SM1.1, right? They simply need to be reviewed and pushed in. ;-)
I think we should release this weekend what we already have finished and reviewed, and see if someone can help along what didn't get ready, so they get in early in 3.7.
When you say "release" this weekend, what do you mean by that? Doug said "go to beta". I assume Doug means to follow the original plan, I quote:
The release date for 3.6 final is August 1st, in keeping with the 4-month release cycle idea, and the first-Friday release date within the month. Beta/Gamma dates haven't been discussed, but I would propose having the gamma release 2 weeks before final, and the beta 4 weeks before the gamma. (If a big problem comes up during gamma, the release would be postponed by 2 weeks.)
AFAICT from all this we are "kindof" doing ok. To me it seems like people would be much happier overall if we could squeeze these in before going to beta - since "beta" implies all new features present:
1. TrueType. I haven't followed the last discussion, but Yoshiki did do a new release? 2. Diego's look style enhancements. Ready to go in? 3. Simulator fixes (someone picked that ball up, right?)
The replacing of the fonts hasn't been much discussed so I assume it could be postponed.
So... if we simply slide the "going to beta" a few days, say... somewhere middle next week, 2nd july? That sounds like it could buy us a few days to wrap up the three above. Doug of course has the last word.
Personally I think it is very important that the community stands behind the release - and there *really* has been pressure about at least TrueType and Diego's stuff so I think we should try our utmost to accomodate that. I really do.
Finally I come to SM1.1. My life is a bit hectic this week. :-) As always of course, but this week is a bit more. We will have our wedding party on saturday and THEN my vacation begins which means I will be able to push on SM1.1 for real.
Personally I think that SM doesn't *really* need to follow the rules of "beta"/"gamma" as strictly because it is in many ways a tool built "on top" of Squeak and not as much BEING Squeak itself. This can of course be discussed. :-)
Anyway, I am committed to it and I really want to deliver SM1.1. I also think it is needed to be deployed *before* we get into 3.7. Why? Because otherwise we will start breaking 3.6 when moving forward with 3.7 - the lack of releases will have that effect.
In practice this means that SM1.1 needs to get in place before we start releasing new versions of the packages that are now being introduced in 3.6 (the ones being a part of official Squeak like for example Scamper, Celeste etc). Exactly what this means to our cycle I am not sure - I can only say that I will be working hard on it and that we will deliver (Brian has done a great job of implementing a proper cache etc), I just can't promise *when*.
Ok, now I hope this post can spur a productive creative discussion instead of people "moaning and bitching" about how the process is moving. ;-) ;-) <- Note plenty of smileys here.
We are all in this together and learning as we go. Personally I am proud of all of us doing such a great job sofar. Marcus, Daniel, Doug and Brent has really gotten the harvesting going for example (sorry if I missed any names here).
As Marcus (I think) said - this is meant to be fun! Let's keep it that.
regards, Göran
Sorry about finally catching up with this now... it's a bit of a struggle to keep up with the lists, and incorporating stuff, and trying to get update-broadcasting to work again. (I haven't made much progress on this last item, but publishing updates by copying files to the ftp site myself is not all that bad, and actually somewhat faster if it's a large batch, so I'll probably just do that for a while.)
On Wednesday, June 25, 2003, at 10:46 AM, goran.krampe@bluefish.se wrote:
...
I think we should release this weekend what we already have finished and reviewed, and see if someone can help along what didn't get ready, so they get in early in 3.7.
When you say "release" this weekend, what do you mean by that? Doug said "go to beta". I assume Doug means to follow the original plan, I quote:
The release date for 3.6 final is August 1st, in keeping with the 4-month release cycle idea, and the first-Friday release date within the month. Beta/Gamma dates haven't been discussed, but I would propose having the gamma release 2 weeks before final, and the beta 4 weeks before the gamma. (If a big problem comes up during gamma, the release would be postponed by 2 weeks.)
AFAICT from all this we are "kindof" doing ok. To me it seems like people would be much happier overall if we could squeeze these in before going to beta - since "beta" implies all new features present:
- TrueType. I haven't followed the last discussion, but Yoshiki did do
a new release? 2. Diego's look style enhancements. Ready to go in? 3. Simulator fixes (someone picked that ball up, right?)
The replacing of the fonts hasn't been much discussed so I assume it could be postponed.
So... if we simply slide the "going to beta" a few days, say... somewhere middle next week, 2nd july? That sounds like it could buy us a few days to wrap up the three above. Doug of course has the last word.
Yes, I'd agree with this as a sort of compromise. We are starting to cheat here a bit by pushing back the beta date but not the final release date, but I don't think we're at a dangerous point yet. So, let's go with July 2nd.
Personally I think it is very important that the community stands behind the release - and there *really* has been pressure about at least TrueType and Diego's stuff so I think we should try our utmost to accomodate that. I really do.
Yes. There was discussion about which has more priority for releases... the Date or Features, and we more or less agreed to give Date priority. I interpret this to mean that Date is most important, but Features are not completely unimportant. In other words, if we've only completed 4 out of 11 features as we approach a beta date, we might consider postponing the date. But if we've completed most features (say, 8 or more), the date stays. We definitely don't want to postpone dates until all features are completed... the list of features is really more of a goal than a strict requirement. For example, at this point I could imagine that waiting to get closures in might delay things by up to several months, if we wanted to completely resolve licensing issues and all other issues.
My example above of when to consider postponing a release is sort of fuzzy... Daniel's description in an earlier post is a bit more objective.
Finally I come to SM1.1. My life is a bit hectic this week. :-) As always of course, but this week is a bit more. We will have our wedding party on saturday and THEN my vacation begins which means I will be able to push on SM1.1 for real.
Personally I think that SM doesn't *really* need to follow the rules of "beta"/"gamma" as strictly because it is in many ways a tool built "on top" of Squeak and not as much BEING Squeak itself. This can of course be discussed. :-)
I've been assuming that SM1.1 wouldn't necessarily be incorporated in the 3.6 final image... just that it would be available by the time 3.6 was released. So it doesn't need to be ready by beta.
Maybe set a goal for yourself to get it done by sometime in mid-July, even if it is a bit buggy. :-)
Anyway, I am committed to it and I really want to deliver SM1.1. I also think it is needed to be deployed *before* we get into 3.7. Why? Because otherwise we will start breaking 3.6 when moving forward with 3.7 - the lack of releases will have that effect.
In practice this means that SM1.1 needs to get in place before we start releasing new versions of the packages that are now being introduced in 3.6 (the ones being a part of official Squeak like for example Scamper, Celeste etc). Exactly what this means to our cycle I am not sure - I can only say that I will be working hard on it and that we will deliver (Brian has done a great job of implementing a proper cache etc), I just can't promise *when*.
Ok, now I hope this post can spur a productive creative discussion instead of people "moaning and bitching" about how the process is moving. ;-) ;-) <- Note plenty of smileys here.
We are all in this together and learning as we go. Personally I am proud of all of us doing such a great job sofar. Marcus, Daniel, Doug and Brent has really gotten the harvesting going for example (sorry if I missed any names here).
As Marcus (I think) said - this is meant to be fun! Let's keep it that.
regards, Göran _______________________________________________ Squeakfoundation mailing list Squeakfoundation@lists.squeakfoundation.org http://lists.squeakfoundation.org/listinfo/squeakfoundation
squeakfoundation@lists.squeakfoundation.org