The Discussion Topic to RFC

[MichaelC]

Alot of users aren't confident enough or for whatever reason such as not enough knowledge on the subject to post an RFC, so they post a Discussion topic about a new feature. There it gets alot of votes and comments but because it was never made into an RFC before the feature freeze deadline, although that feature has had a lot of comments and votes, it still can't be added to that version.

It might make sense to, when the feature freeze deadline comes along, for the development team, to go along the discussion topics and decide whether there is enough content/votes/information in them to make it into an RFC.

By this I mean topics such as here and here which have quite a fair amount of support, but because they were never made into RFCs, therefore they won't be added until the next version, despite having been through a process that is similar to what previous RFC's had been through.

[MartinTruckenbrodt]

Hello,
I've been told by naderman to program a patch which should add (most of) the features of one of my MODs. But I can't find any usefull information how to do it.

Bye Martin

[naderman]

Hello,
I've been told by naderman to program a patch which should add (most of) the features of one of my MODs. But I can't find any usefull information how to do it.

Bye Martin

http://www.phpbb.com/development/get-involved/

[MartinTruckenbrodt]

Hello Nils,
thanks a lot for the link.

Perhaps a sticky topic somewhere at area51 would be a good idea.

Bye Martin

[Oleg]

I will offer my personal opinion here.

It might make sense to, when the feature freeze deadline comes along, for the development team, to go along the discussion topics and decide whether there is enough content/votes/information in them to make it into an RFC.

I disagree, and here is why. Feature progression looks something like this:

1. Discussion - whether a feature makes sense, is possible to implement, would be used, etc.
2. RFC - technical design.
3. Patch in progress - someone is writing code.
4. Merged - code is finished and accepted.

The jump from 3 to 4 is pretty small. Someone who started writing code, and especially if they wrote a good part of it, has already invested time into a feature, and is pretty likely to finish it. The jump from 2 to 3 could be pretty large because writing an RFC is something that takes, let's say, about a day, whereas implementing it may take a week, or weeks. The jump from 1 to 2 is not as large as from 2 to 3 but it's substantial compared to 1 itself - it may take 5 minutes to post a short discussion topic, 15 minutes to post a more comprehensive topic but writing an RFC for just about anything will probably take a day.

A version is declared finished when all features in it are at stage 4. In the steady state I would say that only features in stage 3, where implementation has started, should be included in a future release at the time of feature freeze. Technically this is what was advertised for 3.1 also:

All features which do not have an implementation at that point, will be postponed to 3.2.
- viewtopic.php?f=99&t=33015&p=213231&hil ... ze#p213231

... but, given that 3.1 is the first release under this model, some features in stage 2 were accepted as well. Accepting features in stage 1 would add even more features that have no owner (someone vested in taking them all the way to stage 4), potentially delaying the release forever.

[MichaelC]

You give the impression from your post that the original RFC Poster is the person who has too develop the patch for it. As my understanding goes this is not the case?

Although, generally I can see what your saying that there is a difference between a discussion topic and an RFC. However, I'm not saying introduce every idea discussed. Just 3-4 maximum that have enough content in them to be easily made into an RFC. And I don't suggest 3-4 massive features such as a WYSIWYG Editor. I mean maybe one bigger RFC like a WYSIWYG Editor, but other smaller ones such as the Last Topic Titles discussion topic,the ACP Add User discussion topic. Of which there isn't much work involved in.

[Oleg]

You give the impression from your post that the original RFC Poster is the person who has too develop the patch for it. As my understanding goes this is not the case?

You are correct. Anyone may implement any feature, or submit an RFC for any discussion topic. However, if nobody has done that for a particular topic it suggests to me that there is no one who is both interested in that particular feature and able to move it forward.

By "owner" I meant an "effective owner" - someone who is actively working on getting the feature implemented. It does not have to be the same person throughout the entire life of a feature, nor does it have to be a single person at a time.