[RFC] Release Strategy Following 3.1

[MichaelC]

A possible solution to maintaining 3.0, 3.1 and 3.2 concurrently is to declare 3.1 an interim release of sorts, which will not be maintained long after 3.2 is released.

As a board admin, this worries me a bit. It makes the impression that 3.1 will be a unfinished beta. Who would recommend installing it if that's the case? without even knowing when is 3.2 going to be released.

That was in response to viewtopic.php?p=234789#p234789 and the following post. The issue is not with 3.1's quality, but rather that it may take some time before 3.x + extensions fulfills the needs of 3.0 + modifications boards. There is another post on this somewhere that I cannot find at the moment.

The idea is that boards that do not have many modifications can upgrade to 3.1, and subsequently to the following 3.x releases. Boards that do have modifications will wait for those modifications to be ported to 3.x as extensions. If 3.2 is released before the modifications are so ported we can de-support 3.1 instead of de-supporting 3.0.

How would you decide if enough MODs had been ported to extensions? Considering a lot of MODs don't get updated by their authors anymore you can't say something like half the MODs in the DB, or even a quarter probably.

Apart from that, 6 months would be plenty long enough for those that are going to port MODs to extensions to do it and have their MODs in the Development Forums, 12 months should be long enough to get them validated.

Can anyone remember how long it took to port most phpBB2 MODs to phpBB 3 that did get ported?

[Oleg]

How would you decide if enough MODs had been ported to extensions? Considering a lot of MODs don't get updated by their authors anymore you can't say something like half the MODs in the DB, or even a quarter probably.

Number of people running 3.0 boards?

Apart from that, 6 months would be plenty long enough for those that are going to port MODs to extensions to do it and have their MODs in the Development Forums

Do you have any data supporting this assertion?

[MichaelC]

How would you decide if enough MODs had been ported to extensions? Considering a lot of MODs don't get updated by their authors anymore you can't say something like half the MODs in the DB, or even a quarter probably.

Number of people running 3.0 boards?

That makes no sense, I asked how many MODs being ported would qualify as enough to support 3.1.x
In reply to:

If 3.2 is released before the modifications are so ported

If you mean that all MODs run on boards would need to have been ported before support is dropped, 3.0 will probably still need to be supported in 10 years time...

Apart from that, 6 months would be plenty long enough for those that are going to port MODs to extensions to do it and have their MODs in the Development Forums

Do you have any data supporting this assertion?

Yes, experience.

[callumacrae]

Apart from that, 6 months would be plenty long enough for those that are going to port MODs to extensions to do it and have their MODs in the Development Forums

Do you have any data supporting this assertion?

Yes, experience.

+1, except for crazy people like nickvergessen with a million mods xD

[MichaelC]

Apart from that, 6 months would be plenty long enough for those that are going to port MODs to extensions to do it and have their MODs in the Development Forums

Do you have any data supporting this assertion?

Yes, experience.

+1, except for crazy people like nickvergessen with a million mods xD

Obviously some people might take slightly longer to update who have a number of large mods and not as much time but generally 6 months is plentiful, then obviously it takes a while for validation. To name a few, Rich, Nick, Primehalo (he has literally about 30-50 MODS) but we are talking about the majority.

[Oleg]

https://gist.github.com/8b229d974e1cced63d39

[MichaelC]

https://gist.github.com/8b229d974e1cced63d39

Some expansion on that:
There would be the dev-ascreaus branch and the ledges branch.

When a release is made dev-ascreaus is merged into ledges.

All new ledges go into the ledges branch.

Every couple of months a sub-release is made with a version number like 3.1.0.1 which is identical to the latest maintainace release (using the example 3.1.0.1 it would be identical to 3.1.0) but it includes new ledges but does not include any bug fixes.

Ledges branch is merged into develop-ascreaus and develop.

This is for mods in development primarily that can then require a sub-version rather than waiting 6 months to even get a new mod into alpha. Although ultimately mods in the database could also require a sub-version.

Updating wouldn't be a large issue if a user wanted it for his/her board as it would only include updates for ledges. Also as such it would not require QA and testing so releases could be done instantly without issues.

Thoughts?

[Meis2M]

this is my opinion and i love all of people working hardly on phpBB.

i think number of DEV team member is low. maybe it is a reason for low speed for new releasing of phpBB.
i think 20 member At least must work on phpBB 3.1 directly. some people work only in ACP some people only work on AJAX system and...
I think it is better to have a larger circle of DEV team. we have many Intelligent users on phpbb.com that can help for develop phpBB.

i know we have Many Masters in phpBB but number of them is low.
Nils Aderman is dev leader he is one of the best Moderator of phpBB.
callumacrae Master of AJax system in 3.1
Dear DavidIQ,igorw,nickvergessen,Oleg and................... all of phpBB team. Thanks a lot from yours.

[psoTFX]

The development team has always been historically "small" - but these days it's at its largest "ever". Anyone can contribute code, etc. And to be honest team size isn't "really" the issue (I use quotes intentionally). The proposal to implement specific (unwavering) merge windows and point (alpha/beta) release based on them is the way forward imho.

[EXreaction]

Like psoTFX said, you don't need to be on the development team to write code in phpBB.

I would actually lean the opposite way and have less people on the development team (just a few focused on merging patches and leading development of specific releases). While I know there is some push (and benefits) to be on the teams if you really support the project, I would think that keeping more people outside of the teams who just create patches and develop would be helpful because it would force a bit more openness and it might make people feel like they can actually write code for phpBB more easily (that they don't actually have to be on the development team to do so). This is of course just my opinion, but either way, with the much more open form of development available now, you don't need to be on the teams to develop phpBB.