Talk:Summer of Code 2006/ideas

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search

Teh 3vil[edit]

Wikimedia should not participate on this. We all working on open source projects. Open for ALL. But Google say no to Cuba, Syria and North-Korea. What do we have do do with politics? Nothing! We are free. We shouldn't support the ideas of the US-Goverment like Google does. It should be a free thing for all - and not for some. What google does is not to handle with the Wikimedia-idea! Kenwilliams 08:26, 15 April 2006 (UTC)Reply[reply]

Yeah, I sure *hate* the fact that we have to *follow the law* of the country we're in! That's tyranny! --brion 08:41, 15 April 2006 (UTC)Reply[reply]
Wikimedia Foundation Inc. should be under the same restrictions as any other organisation in the US. That means that Brion's job could not be offshored to Pyong-Yang. Yet. Ken, I think you are wrong. If google pays a student 4500 USD and the code gets done, it can be downloaded even in countries like Cuba if their authorities aren't restricting it. -- Mathias Schindler 09:39, 15 April 2006 (UTC)Reply[reply]

What's "better" - a bad dictature or a democracy that forget his democratical structures? And it's clear, that google does ist this way. But we at the Wikipedia shouldn't support this. The Main Wikipedia directive is FREE FOR ALL. So we should only work together with people who see it in the same way. Kenwilliams 18:28, 21 April 2006 (UTC)Reply[reply]

Then please cease working with Wikipedia immediately! Wikipedia is operated by the Wikimedia Foundation, Inc. Like Google it's an American company and subject to United States federal laws. --brion 06:28, 25 April 2006 (UTC)Reply[reply]
Even if it weren't Google, by US law Wikimedia couldn't get any Cubans, Syrians, North Koreans, Taliban students, or whatever to work on this because of trade law. If you have a problem with the Wikimedia foundation following US law (which is, in some sense, a valid problem -- embargos don't go well with open source), find a wiki encyclopedia on a darknet. --Geoffrey 05:14, 3 May 2006 (UTC)Reply[reply]


Is there a way to say "HELL NO!" to WYSIWYG wiki editing? Jon Harald Søby 09:07, 15 April 2006 (UTC)Reply[reply]

You could save your "Hell no" until the question arises to activate this feature by default on a Wiki you participate in. -- Mathias Schindler 09:39, 15 April 2006 (UTC)Reply[reply]
This may seem inappropriate, but actually has a nice WYSIWYG editor. Of course, the offer both possibilities based on user preferences. It would definitely lift the learning bar required to edit on wikipedia and the other wikis. Ansell 23:30, 20 April 2006 (UTC)Reply[reply]
Additionally, can we say "HELL NO!" to AJAX? The potential problems scare me. Keep it simple.
This one seems to have dissapeared when the list got reordered, is is something that shouldn't be a proposal anymore? Henna 14:09, 25 April 2006 (UTC)Reply[reply]

How do we vote, modify etc...[edit]

So I really like the AJAX and WYSIWYG proposals. Am I supposed to write "*support"? :) Or, where do we flesh out these proposals? They're obviously very vague and open-ended at the moment. Vandalism heuristics would be great too. Alternatively, is there an option where the person just joins the development team and fixes bugs for 2 months? That would be helpful too :) Stevage 10:33, 15 April 2006 (UTC)Reply[reply]

You can vote all you want, but that's not going to do much. There are really going to be two votes that count: one is the student who picks one of these ideas to work on, and the other is going to be either brion or Tim's OK at the end of the summer when (hopefully) the project is complete and it's time to merge it into MediaWiki or (better) release it as a MediaWiki extension.
As mentioned above, the question of whether this code gets written, and whether it gets installed and activated on a wiki near you, are kind of orthogonal. I think the lead developers of MediaWiki, and most if not all of the development team, encourage that new features can be turned on or off by sysadmins easily (either built as extensions or with global enable/disable flags).
Probably the best thing to do right now, if you really support any of these ideas, is find the page on meta that expands on the idea (see WYSIWYG editor, for example) or start one, if it's not started. The more information we can give on an idea, the more likely a student is going to pick it and do it correctly. Bloating this page with lots of details on each idea is probably not the best way to stimulate interest.
And, of course, if you have an idea that's not on the list, just start it. --Evan 14:57, 15 April 2006 (UTC)Reply[reply]

Student : BuThAnE[edit]

Hi, i'm a french sudient (sorry for all mistake i'll do) interrested in the Export project. I would like to know the mentor who will give guidance throughout the duration of the program and talk to him? :)

I am also interested in the Export project, and my question is, does Wikimedia have any requirements for the programming languages and such for backend projects? For example, will Wikimedia only take PHP or other applications in the form of web pages, or would they also accept applications being worked on in Python or such.
If you're going to tie directly into MediaWiki, the only languages that really make sense are PHP, C, and C++. (C/C++ can be used to write PHP extensions.)
If you're going to communicate with MediaWiki code only indirectly, then other languages are fine. Some notes offhand (my personal preferences);
  • Python is great (but mind performance issues)
  • Perl and Ruby should be similarly ok. (but Python is nicer ;)
  • Java is great, if you want to trade some pain for more speed (but mind compatibility with GNU Classpath / GCJ; avoid tying yourself to Sun's proprietary offering)
  • C# is similarly fine (Mono compatibility is a must; tying to Microsoft's proprietary offering is even worse :)
  • C/C++ are fine too but mind safety! We hates buffer overflows. and portability is a must. Code should compile cleanly on multiple *nix platforms and if possible win32. C/C++ code needs to be able to run on both 32-bit and 64-bit platforms, either endianness.
More obscure languages could be accepted if suitable. --brion 06:37, 25 April 2006 (UTC)Reply[reply]
If you're looking for a fast, portable, free, nice language, there's Haskell. DavidMcCabe 22:30, 29 April 2006 (UTC)Reply[reply]

Client API[edit]

There is a message under Client API saying work has started. This interface seems more like a bot query interface and provides no methods for updating records. Isn't something more robust and standardized, such as SOAP or XML-RPC wanted? --IndyGreg 23:34, 24 April 2006 (UTC)Reply[reply]

I'm a student wanting to apply to Google Summer of Code 2006 and so far the MediaWiki Client API is by far my favourite project to apply for! I'm interested in writing a SOAP/REST/XML-RPC API which would allow updating articles etc. as well as simply retrieving them. I don't think the query interface linked to is really an API in this sense, but I believe the code is only at an Alpha stage so perhaps that code could even be integrated into an API at some point.
What does exist is a third party API run by Ontok, I assume they have no direct affliation with the Wikimedia Foundation?
Do you think the API would be better written in PHP or C? I could do either but I have a preference for PHP. Any advice from the wikipedia community for us people applying for this project on what to say in our application? I really don't want to miss out on this opportunity :)
Ben Francis
--Tola 19:02, 25 April 2006 (UTC)Reply[reply]

SoC Student: Permissions and Statistics[edit]

I'm currently interested in implementing the permissions and/or statistics ideas for the SoC, but the ideas themselves are rather vague in details and I am not quite sure where I might find more information or a place to get started. Can anyone point me to other resources that might elaborate on these ideas? Thanks. --AbyssNOLF 22:42, 25 April 2006 (UTC)Reply[reply]

Poke your head into our developers' mailing list perhaps. :) --brion 02:40, 26 April 2006 (UTC)Reply[reply]

Backend: Message System[edit]

There already exists a very lightweight and effective Chat Notification extension to Wikimedia. It currently requires a special open-source multi-protocol Server called because other protocols do not permit event submission in such a lightweight way like PSYC does (You have to deal with a login procedure at least etc). But that's no problem. Once installed psyced acts as an IRC and Jabber server so all you need to do is join the MUC or channel where your notifications are directed to and you are in the loop. It has become a very powerful feature in our PSYC developer chatroom where people would immediately comment on things we just edited. In combination with Jabber transports you should find a way to even send such notices to ICQ, but we haven't looked into that. This may not be a very common solution, so if you're experiencing any problems getting such a solution to run please tempestively consult us. Then again, if this type of solution is not what you are looking for by Message System please specify how it should be implemented differently. Would you want Wikimedia to implement its own builtin multi-protocol server or something like that? Good luck! --lynX 17:24, 29 April 2006 (UTC)Reply[reply]


Greetings, all.

I am thinking about proposing a LiquidThreads implementation. I have two questions:

  1. Is this something people are actually interested in, or is it one of the ones to be 'parred down'? Please say so soon in the later case!
  2. I think I could write a prototype of this in a couple of days. The prototype would not be integrated into Mediawiki; rewriting it as part of Mediawiki would be the summer project. Would this be worthwhile? Would it help you evaluate the proposal?

Thanks! DavidMcCabe 22:11, 29 April 2006 (UTC)Reply[reply]


Hi, I'm interested in working on the XHTML2Wiki Summer of Code idea. Is there someone who is willing to mentor me with this project? Has anyone done anything similar before? How would this differ from the solutions here and would it be OK to have it only import well formed XHTML to start with and then add on HTML 4 and generic HTML support later on? --Yi Ding 22:57, 2 May 2006 (UTC)Reply[reply]


I'm not eligible for the Summer of Code because of age, but I am going to be a college student next year (at MIT) and I do know how to code. Is there a way for me to participate in this for fun (not for pay), or will the IP unsureness for minors still get in the way? --en:Geoffrey 05:10, 3 May 2006 (UTC)Reply[reply]

You're always welcome to hop in and contribute code on your own. You won't get a Google T-shirt out of it, though. ;) --brion 22:48, 3 May 2006 (UTC)Reply[reply]

Heuristics for vandalism[edit]

I'm interested in working on this, but I'm new to the mediawiki codebase and am unsure where the existing code, the IRC notification module, is. Could someone in the know point out where in the tree this is? I'd like to get a feel for everything involved in this.

On a different note, was this project put up with the intention of using a stock bayesian filtering tool, or having us roll our own? What draws me to this one in particular, the the chance to learn about bayes nets.

thanks Catphive

Idea not listed[edit]

I'm a student interested in participating in the Summer of Code. Since I've been editing Wikipedia for quite some time, I've got my own pet idea on how to expand the current wikimedia mechanism, which I would like to implement.

The question is:

  1. Should it be entered on this page for evaluation, or
  2. I should not edit the page, it's strictly for the current developers' use to express their needs, or
  3. It doesn't matter either way, what will really matter is the quality of a proposal, or
  4. It doesn't matter either way, the set of possible projects is already fixed and I'd better spend my time polishing my other proposals? ;)
I'm student too. You can make your proposal both on that page and via Google SoC. But if you really want to take part in SoC you have to submit your proposal at their site. That action just gives you a chance but doesn't prove your proposal acception. For further information reed FAQ and Summer-Discuss-2006 group.

Adding pengo/ite[edit]

I'd like to add a suggestion for a SVG templating engine to the list. The idea is described in some detail here: w:User:Pengo/ite. I can't tell how official or final this list is or who's meant to be adding to it so I'm just adding my suggestion here on the talk page.. Pengo 17:22, 9 May 2006 (UTC)Reply[reply]