is "fork" a 4-letter word?

[code reader]

@SamG:
I wouldn't put it at the "ax to grind" level. i do think it can be an interesting experiment.
i would also say that the phpbb group can determine the results of the experiment, at least in one direction: they can't make it work, but they can prevent it from working.
"open development" requires open communication; typically this open communication is implemented either by a mailing list, or with a discussion forum. since the subject matter is phpbb, obviously only a forum makes sense.
personally, i have absolutely no intention of opening a "competing" bbs somewhere to discuss this open development, and have no intention to log on to someone else's "competing" forum. i expect this is true for most developers.
moving the communication elsewhere is exactly what the "bad fork" is about: splintering the community. this is absolutely not the intention.
so the whole experiment only stand a chance of producing any positive outcome if the phpbb group will allow developers to discuss their ideas, changes and patches here (or over at phpbb.com).

so, as i said, they can't make it work, but they can, if they want to, make it not work.

i have good ground to expect they will allow this to happen: after all, there are the "modders discussion forums" back at phpbb.com, which is, after all, "parallel development".

of course, it can fail all on its own even with full support from the phpbb group: so far i haven't heard from a single developer expressing any interest, and my own contributions might not be of any magnificent scale...


@david:
i absolutely agree with everything you say. this is exactly my point, after all: in the current state of affairs, it is virtually impossible for anyone without "commit" permission to contribute anything of interest: first of all, it is harder than need be to work on something serious with no SCM system. and if i set up my own private scm to track my progress (i did exactly that when i needed to make a significant change to osCommerce), it becomes a major pain to stay synchronized to the main project.
but even if one passes these hurdles, the basic problem still remains: how can i hand my changes for merging upstream?
as you note, creating one huge turd of a patch (or "mod") is not going to fly. anything significant that comes as a single blob just represents too much work for the core team to review.
maybe an open scm will not solve this problem, but imho there is at least a chance it will.
it will allow everyone to view progress, you will be able to comment on bad or non-conforming code and style as they are introduced (things like: "we have an object that does exactly that. please use it") , and as such allow the originator(s) to fix their code and stay conformant during the development, hopefully improving the chance of upstream merge.

in short: i don't know if "open development" will produce anything that you'll eventually merge upstream, but it is quite clear that without it community contribution is almost impossible.

i don't remember a single project that came out of the Community coding project.
and i can't remember a single MOD being integrated upstream. not even something like the attachment mod, a sorely missing feature in phpbb2, was integrated, and this while the author was the maintainer of phpbb2! there are many other examples of pieces of code which would have made the project better ("tweaks for large forums" comes to mind) but just can't be merged upstream because they come in the form of a huge turd.

i think that with open development the picture might have been different.

to summarize: exactly because of what you say i believe that open development is the only way community members can contribute to the core.

even non-core parts, such as styles, translations and even MODs can benefit from open development.


as to your last point: i did not and do not advocate the phpbb team moving to any new tools or changing their work flow.

[SamG]

personally, i have absolutely no intention of opening a "competing" bbs somewhere to discuss this open development, and have no intention to log on to someone else's "competing" forum. i expect this is true for most developers.

It is true for most content developers. Those that become discontent for whatever reason (want to monetize the project, dislike project management/development decisions, like spaces instead of tabs...) can march off with code (perhaps bearing their copyright) and open up a respectable shop elsewhere. It may not be seen as a negative fork, in those cases, especially if people are packing their own code. In other words, your intentions, however well intended, can't constrain open development to the original project anymore than can the GPL. (If you're interested, look up Quanta Plus and Quanta Gold as an illustration of what an amicable fork might look like.) The smaller the scope of the project, the better the chances that somewhere down the road an open development project will undergo an amicable fork, IMHO.

That kind of risk isn't necessarily a show stopper for open development as a concept. It might be possible to reduce that risk through various up front agreements. I'm actually more interested in the risk of project dilution as developers, through an organic process, leverage open development into open project management. Open development means some level of contribution parity if it has any teeth and any scale at all. (Else we're effectively at the Junior Developer concept.) I think the risks to product quality and continuity increase at that point, especially within a single product.

But, that all remains to be seen.

[code reader]

please allow me to repeat the question that may have been drowned in way-to-long response(s):

is it ok with the dev team if i publish the git-repo on the mod developers forum on phpbb.com?

[psoTFX]

The kind of "organic" development being suggested is interesting - but not necessarily appropriate. You cite as an example the lack of inclusion of the attachment mod or any mod. Well that's not quite correct - I back ported a phpBB3 feature to phpBB2, the original captcha code. The simple fact was/is that phpBB2 development was ended, it was maintained but not extended for anything bar security or similar reasons. That was a design decision - sure - the (very) delayed release of phpBB3 wasn't anticipated. But those issues have since and are still being addressed - with smaller feature sets for forthcoming versions, additional development resources, etc.

Open development only works imho when you have a game-plan and someone to regulate it, Linux was given as an example, a good example. It's not merely a case of approving or removing submitted patches - it's about being able to explain and justify why patch A was admitted but patch B was refused. I wouldn't want Linus' job for all the tea in China, look at even some recent "decisions" he's taken and the subsequent backlash it has incurred. People can be very passionate about their ideas and more so their code - take it from someone who had to turn down ideas for several years because they simply didn't fit in with the plan

For a forum system such as phpBB there are literally dozens and dozens of "features" you could include - we know - we have trackers full of them But you need to decide just what "kind" of forum system you're looking to produce. In a practical sense phpBB could be forked into a dozen or more projects all focused on a different path, be it individual or group collaboration, simple discusion, file sharing, team spaces, schedulers, etc. - look at WebEx, Sharepoint, etc. for examples. As an aside these actually interest me more than opening development on phpBB "itself". Ignoring that imho phpBB is already doing much of what you suggest, more so than ever before with yet more changes set to be introduced. With a team of peeps overseeing the current design plan and implementing it - with where appropriate and possible contributions gathered from the wider community, new coding projects and development opportunities.

Wrt the community coding projects - you're overlooking the fact that developers such as Nils and David were recruited from the CCP's. The fact that some of the projects have, yet, to see the light of day doesn't detract from that. They proved useful but not necessarily in the form that was intended by myself.

Wrt the large forum topic - that was a great topic. What's more it directly fed back into phpBB development with various changes being derived iirc (it's been a while lol ... but I recall myself and Bart and others discussing various points raised at various times). But again this is another example of a potential "specialist case" - where you can easily get bogged down in just what you're trying to achieve.

[code reader]

first, i need to say that the lack of any response from any potential "community developer" means, to me, that the "open development" model i was talking about is probably, as you say, not appropriate to this project.

however, i would still like to relate to some of the points you made:

You cite as an example the lack of inclusion of the attachment mod or any mod. Well that's not quite correct - I back ported a phpBB3 feature to phpBB2, the original captcha code.

i don't see how backporting an internally (i.e., team) developed feature from phpbb3 to phpbb2 is relevant to a discussion about community contributions.

Open development only works imho when you have a game-plan and someone to regulate it, Linux was given as an example, a good example. It's not merely a case of approving or removing submitted patches - it's about being able to explain and justify why patch A was admitted but patch B was refused.

actually, it's about much more than that: it's about giving the patch developer a stage to explain why the patch is required in the first place, a stage to show the patch for review by community and "team" alike, providing an iterative process to refine the patch until it meets standards, and eventually, non-arbitrary (though not "democratic", either) decision of inclusion or rejection.

linux may be a wrong place to look for reference: many in the community are actually individuals who are paid by their employers to develop pieces they want included. nothing is wrong with that, but the level of professionalism and amount of energy these people can put into their "community participation" in the game is on a different level from phpbb community.

more relevant reference point would be smaller projects, where majority of people, maybe even the chief developer have a "day job", but still able to have a lively community participation. such projects do exist.

Wrt the community coding projects - you're overlooking the fact that developers such as Nils and David were recruited from the CCP's. The fact that some of the projects have, yet, to see the light of day doesn't detract from that. They proved useful but not necessarily in the form that was intended by myself.

i am not overlooking this fact. however, the "ccp" was never presented as an "entry exam" to the team, and the fact is that from the 10 or so proposed projects, not a single one actually made it to the code. 2 years after launch it doesn't seem likely that any of them is still alive.
it is my contention that a major factor in this failure is the fact that all the work on these projects was done in the dark. i believe that doing the same thing in public would have worked better.

personally, i wanted to try and contribute by fixing the now-broken search mechanism, but there is no way with the current mode of work this is possible: if i contribute one huge patch that touches several files, there is no way, (as david noted) that this patch would even be considered, and there is no venue for me to present a series of more manageable size patches, and there is no mechanism for these patches to be reviewed, commented upon, improved, tested, and eventually included.
the result is that the search is likely to stay broken indefinitely.
oh well....

[naderman]

i am not overlooking this fact. however, the "ccp" was never presented as an "entry exam" to the team, and the fact is that from the 10 or so proposed projects, not a single one actually made it to the code. 2 years after launch it doesn't seem likely that any of them is still alive.
it is my contention that a major factor in this failure is the fact that all the work on these projects was done in the dark. i believe that doing the same thing in public would have worked better.

We agree with that, and that's why we're going to have another try this year with a much more open process (code in open repositories, regular public reports on progress, ...)