Note: If there is anyone reading this who knows about how many public access keyboards we plan for the Maker Faire booth this year, please let me know. Stephen: I really need your help with the bylaws as soon as possible and bookmark handouts later in the week. I also want to encourage all the GSoC student applicants who weren't accepted to apply for a self-study grant at http://wiki.creativecommons.org/Special:AddData/Grant_Application to do the work you proposed over the summer. For the benefit of the chapter, I have included a list of some of the proposed software projects below. Thank you!
Johan, it's possible that less than half the Foundation engineers knew squid logs are aggregated on the toolserver. You and Laszlo (and possibly other student volunteers) are both way ahead of schedule, even though you weren't accepted into the program. This should not be a problem because I will see that you are reimbursed for your good and helpful effort. According to the instructions on Meta and from the Foundation, the correct way to do that is to form a chapter, so that's what I'm working on. That and handouts for Maker Faire, which may or may not have keyboards and displays this year.
Good news, Meadowlark was accepted into Berkeley, and he might need to take a transfer-in class over the summer, so that means I'm going to try to do his project. If you or Laszlo as non-monoglots can help and check my and Meadowlark's work, that would help a lot.
My first question is how do you feel about the plan: How do you feel about modifying the Mediawiki Quiz extension -- http://www.mediawiki.org/wiki/Extension:Quiz -- with regular expressions upon reading from a Gift: namespace (do you recommend using the ArticleViewHook hook?) to turn things like...
:: Question 2 :: What's between orange and green in the spectrum? {=yellow # correct! ~red # wrong, it's yellow ~blue # wrong, it's yellow}
...into...
{'''Question 2''': What's between orange and green in the spectrum? |type="()"} + yellow || correct! - red || wrong, it's yellow - blue || wrong, it's yellow }
...in the default namespace for quizzes? I've been looking through things like...
http://www.php.net/manual/en/reference.pcre.pattern.modifiers.php http://www.php.net/manual/en/reference.pcre.pattern.syntax.php http://www.php.net/manual/en/pcre.examples.php http://www.php.net/manual/en/intro.pcre.php http://www.php.net/manual/en/language.types.string.php
... so I realize the properties of the task require me to offer it to a non-monoglot such as yourself if you want to attempt it.
Bad news, I didn't get in under the deadline yesterday for adding all of the subcategory criteria in http://meta.wikimedia.org/wiki/Talk:Wikimedia_California/projects#Budget_bre... (it and the following sections like success criteria have comments) for the things on http://meta.wikimedia.org/wiki/Wikimedia_California/projects
Good news, we can also apply to fund those projects though applications at http://wiki.creativecommons.org/Special:AddData/Grant_Application
So I have four action items now: GIFT conversion (unless Johan or Laszlo or someone else wants it), bylaws for wikiBusinessMeetings (Stephen promised to help with the bylaws), grant application success criteria, and MakerFairePedia handout instructions and forms (Stephen was going to help copy the bookmark handouts on cardstock and slice them up at Fedex Kinkos.) Since you're way ahead of schedule even though you weren't accepted, if you could help with GIFT conversion now that would be awesome, and I will do what it takes to make sure you get recognition and rewarded.
Best regards, James Salsman
On Sun, May 16, 2010 at 5:53 AM, Johan Gunnarsson johan.gunnarsson@gmail.com wrote:
Interesting to hear. Did they elaborate exactly what part they believed was "impossible"?
Speaking of Wikitrends, I'm test running the new version on toolserver now. It does monthly and weekly trends too (which imo are far more interesting.) The total input size 400-500GB uncompressed, but thanks to some clever filtering and caching a single update takes less than 5 minutes.
Link: http://toolserver.org/~johang/wikitrends/english-uptrends-this-week.html
Feel free to spread the link around.
On Thu, May 13, 2010 at 18:29, James Salsman jsalsman@gmail.com wrote:
I have asked the Chapters Committee, for those Chapters who are membership organizations, whether they are open to members beyond the geographical designation of the Chapter -- I believe the Chapters Agreement and process documents indicate that they must be.
On Wed, May 12, 2010 at 5:14 PM, Johan Gunnarsson johan.gunnarsson@gmail.com wrote:
Sorry for the late reply. Dunno really what this (I don't live in California?) is but I want to say that I'm still interested in my project (Popular related articles) but I'm busy atm.
Johan, on a personal note, your case was particularly disturbing to me during the Google Summer of Code selection process, because multiple mentors said that what you have already accomplished -- with WikiTrends at http://users.student.lth.se/dt05jg2/wikitrends/en/24h.html for example -- was impossible, or harder than you could possibly realize. You were clearly scored lower than you should have been because of this mistake. Oddly, three members of the Foundation's engineering and technical staff repeated the same mistake at a talk they gave at Xerox PARC last week. I will do what it takes to see that this mistake is corrected.
Best regards, James Salsman
--- example proposed California Chapter development tasks ---
http://socghop.appspot.com/gsoc/student_proposal/review/google/gsoc2010/maci...
study and implement co-process communication between Mediawiki and RTMP server (microphone audio upload from flash)
Why it's likely to succeed: Maciej is a Red5 expert (I found him on the Red5 mailing list) and within a day took the time to understand the basics of the Mediawiki file upload API. He's capable with both Red5 (Java) and rtmpd (C++) Flash servers, and without prompting immediately understood the details about gathering the upload from a PHP Mediawiki hook, converting it to ogg vorbis, and sending it to Mediawiki. I doubt this will amount to a lot of work.
Why it's worth it: This would be the first Mediawiki extension with a co-process, and it's likely to be useful to figure out the right way to bundle that while the work on the extension manager is still going on. Plus, it will enable easy pronunciation uploads for all of the Wiktionaries, some of which could really use it. That's a direct user and editor advantage. Did you know that en.wiktionary has audio for "Hola" but es.wiktionary does not?
http://socghop.appspot.com/gsoc/student_proposal/review/google/gsoc2010/deep...
"GIFT Conversion" (for Quiz extension)
full version: http://en.wikiversity.org/wiki/Google_Summer_of_Code_2010_Proposal:_GIFT_Con...
Why it's likely to succeed: I've already met with Meadowlark twice in person and several times online, discussing the architecture in detail. He's an accomplished programmer, and while he's new to PHP, he understands exactly what needs to be done and is highly motivated and committed to accomplishing it. He's picked the minimum effort necessary to enable Quiz extension service of GIFT quizzes, which can be
Why it's worth it: 98% of the Creative Commons-licensed quiz content from UK's Open University, which represents 5,000 hours of coursework, can be exported as GIFT multiple choice questions and used directly in Wikiversity content. What other effort could provide as much legitimacy and utility to Wikiversity?
http://socghop.appspot.com/gsoc/student_proposal/review/google/gsoc2010/lkoz...
Releasing WikipediaVision as an open source package and further visualization and data mining ideas
Why it's likely to succeed: The initial part of this effort is already done, but unfortunately I think people got an impression that it was only releasing existing work before Laszlo had a chance to flesh out the proposal with the other aspects he included, involving automated vandalism flagging (within the existing Abuse Filter/Edit Filter framework) and other visualizations. Laszlo is a very advanced PHP programmer who has already proven that he can make useful Mediawiki extensions and it would be a shame to waste his talent.
Why it's worth it: The initial release of WikipediaVision has the potential to be a decent tool, and with the proposed enhancement's I suspect it would be a valuable anti-vandal (and vandal-fighter recruitment) tool. The other projects Laszlo proposes stand on their own merit and I urge people, especially those of you who might have dismissed it early on, to read all of the investigations and efforts which Laszlo has since proposed.
http://socghop.appspot.com/gsoc/student_proposal/review/google/gsoc2010/joha...
Most popular related articles
full version: http://www.mediawiki.org/wiki/User:Johang/Most_popular_related_articles
Why it's likely to succeed: Johan has already been a Mediawiki GSoC student, so he has that experience. He's also the author of "Wikitrends" -- http://users.student.lth.se/dt05jg2/wikitrends/en/24h.html -- so in fact he has already done the main part of the work of agregating the squid logs. I am sure that those of you who ranked this down saying it was too large of an undertaking must have missed that. Could there be anyone more qualified for this project than the author of Wikitrends who is already doing the statistical work involved? The only remaining portion is to make a transcludable list of related articles for each article's sidebar, and that only involves choosing between several existing related article generation algorithms, and simply sorting them by their tally in Johan's existing statistics tables.
Why it's worth it: The motivation for this extension is this graph from the strategy group's surveys: http://strategy.wikimedia.org/wiki/File:091207_QOTW.png -- Adding this simple statistical information to the sidebar should help contributors with those two largest hurdles shown in that graph.
http://socghop.appspot.com/gsoc/student_proposal/review/google/gsoc2010/cvil...
Confidence Assessment within Computer Adaptative Learning Environment
full version: http://www.mediawiki.org/wiki/User:Cam_Vilay/Proposal
Why it's likely to succeed: This is a simple user experience protocol analysis study. The only coding involved is making different mock-ups of the various ways learner adaptivity in the Quiz extension could measure the confidence of learner responses. The main effort is videotaping kids and beginners to see how they react to the various choices.
Why it's worth it: How many extensions to we have that provide a valuable feature but aren't used because they aren't well designed? I think it's important to show that we're capable of doing a bona fide user experience evaluation for the kind of further enhancement of the Quiz extension which would help make all that additional quiz content in Wikiversity more useful.