[TYPO3-v4] Motivation for making patches or how to speed up the review process

Dmitry Dulepov dmitry.dulepov at gmail.com
Mon Apr 12 10:41:54 CEST 2010


Hi!

Below is my PERSONAL views about it. They are not necessarily correct but I came to them observing core development for over 5 years.

Jigal van Hemert wrote:
> It's very demotivating if someone takes the trouble of writing a patch
> for an existing problem or a new feature and the RFC is not reviewed.

True.

> It's very demotivating if that person sees other RFCs get votes within
> hours or days after posting it to the core list.

It depends on RFC. One RFC can be very simple (like quotes around attributes), the other can change behavior of IRRE, which involves TCEmain changes, etc. That is much harder to test. It may require hours, sometimes days to find and test all cases. Often it requires creation of a special environment and its reset after each test round. This takes tons of time!

Add here a question of motivation. What would motivate developers to do such testing for hours? Don't tell me about romantics of open source. It is not going to work :)

> Some time ago I heard strong rumours that the Association would hire
> someone to spend at least some time on reviewing the forgotten RFCs or
> at least getting someone to review those lingering patches. Until now
> nothing seems to happen.

At least half a year ago official position was "we do not pay for development". I heard this from two of the top TYPO3 persons independently. Strangely, these rules does not apply to v5 team.

> Ideas for solutions which might help:
> -------------------------------------
> 
> 1. As suggested by Susanne, provide a step-by-step explanation how to
> reproduce the problem and how to test the solution.

Doesn't help. I tried. Even a special extension that reproduces a problem does not always enough.

> 2. Provide unit-tests so it becomes easier to test and review (Oliver
> did a presentation at T3CON09 [1], although the real How-to is missing
> from the slides)

Hard to create, extra effort for the developer, no real motivation.

> 3. A "REMINDER" should trigger the core devs to review an RFC or at
> least respond with remarks on what's missing, etc.

Not working, we do that regularly.

> 4. A "REMINDER #2" (or higher) should trigger someone higher (release
> manager?) to take action. This could be assigning a core dev to review
> it or ...

Will not work. Nobody can ~assign~ a task to coredevs because they are not paid. Tasks can be assigned only to people who are employed. Coredevs alreayd do lots of work in their spare time. They are free to develop when they want and what they want.

> 5. ...<insert your ideas here />...

Only one that really works: have at least two employed high quality developers, who will do it daily.

Core team already tried paid model (paying for important features) and it worked well. I strongly believe TYPO3 must have paid developers in order to have a quick development process.

-- 
Dmitry Dulepov
TYPO3 expert / TYPO3 security team member
Read more @ http://dmitry-dulepov.com/


More information about the TYPO3-project-v4 mailing list