On 02/15/2012 04:39 PM, Andrew Garrett wrote:
On Wed, Feb 15, 2012 at 4:29 AM, Jeremias Nunez jeremias.np@gmail.comwrote:
Hey everyone,
Reading around MW and looking into some ideas for GSoC I ran into this bug, https://bugzilla.wikimedia.org/show_bug.cgi?id=14950, which I noticed (correct me if i'm wrong) isn't currently being worked on, and as Roan Kattouw said on one of the last comments, hasn't ever been written yet (i know it's kind of old, but anyway).
I wanted to get some feedback on the next couple of things:
1 - Is this a feature that could actually be needed/helpful/useful? 2 - Is it true that it's currently not being worked on? 3 - Would it be too big of a project for a single person to take on? (given the amount of time to actually implement it for GSoC)
We discussed this on IRC. A quick recap:
I tried to implement this when I completely refactored the preferences system in 2009. It was eventually reverted in r49932http://www.mediawiki.org/wiki/Special:Code/MediaWiki/49932. The main blocker was basically considering a way to decide *which* preferences would have their values synchronised. A UI would need to be developed for that and you'd need some extensive consultation on that fact.
If you were to implement this, you could potentially use my original implementation as a guide, though it is reasonably "in the guts" of MediaWiki so you'd have to be reasonably confident "code diving" into unfamiliar software packages.
I'll defer to Sumana on whether or not the project is of an appropriate size for GSoC.
—Andrew
Jeremias:
First, some general advice: look at previous people's experiences. Past projects:
http://www.mediawiki.org/wiki/Summer_of_Code_Past_Projects
And, just as a reminder: You're trying to budget three months of working about 40 hours per week, plus some ramping-up time in the spring and whatever you can fit around your other commitments between now and late April. The timeline:
http://www.google-melange.com/gsoc/events/google/gsoc2012
Now, specifically about Global User Preferences. From what Andrew has said about this, it's a task that would just be a few hundred lines of code, perhaps two weeks' worth of work, for a skilled MediaWiki developer with years of experience, comprehensive designs and specifications. I imagine that if you either had the UI specification OR years of experience, but not both, the work would take somewhat longer. But you don't have either, so maybe you should spend a few days investigating the issue to figure out how big and unwieldy it is, on a code level and on an interaction design level. Then you can think about what the majority of the work would be, and perhaps consider whether you want to do THIS as a project, or whether it gives you an idea for a related project that is less unwieldy.
I asked Andrew for examples of past MediaWiki GSoC projects that were well-scoped, and he suggested Visual Diff, Image Rendering Daemon, and the SocialProfile work. So you could look at those.
Looking forward to continuing our conversation as you learn more and think about what a well-scoped project would be.
You wrote:
When you say "extensive consultation" are you referring to the UI or to which preferences could be made global? I think we talked about this on IRC too but I don't quite remember.
It will be worth your time to install and learn a real IRC client, such as xchat, that saves IRC logs. You can also read IRC channel logs at http://toolserver.org/~mwbot/ for #mediawiki and http://prototype.wikimedia.org/logs/%23wikimedia-dev/ for #wikimedia-dev.