Last modified: 2014-11-17 10:35:48 UTC
I think there is a need for a user-generated help system like for instance help.openstreetmap.org is. The rules of English Wikipedia are a big barier in acquiring of new users and editors. New users are not familiar with wiki interface and there is no "easy" place where they could ask something (if they are not at least so tech savvy to leave a note on a discussion page, usualy without a signature). Opening an easy interface for help will be a big benefit for English Wikipedia. Incorporate the Q&A system in SUL. Open source software should be used (OSQA, Pligg etc.). The HELP site over at OpenStreetMap is a big success so far I think.
What software should we use for the Q&A system? It sounds like you want it to be developed first, in which case this shell request should be closed LATER.
(In reply to comment #1) > What software should we use for the Q&A system? It sounds like you want it to > be developed first, in which case this shell request should be closed LATER. Some software options: - http://www.osqa.net/ (Python) - http://www.question2answer.org/ (PHP 4.3+, MySQL 5) - http://shapado.com/ (Ruby) - http://www.lampcms.com/ - http://pligg.com/ - http://askbot.org
And I think the help.en thing should just be like a generic redirect, so we don't need one installation per wiki etc...
(In reply to comment #3) > And I think the help.en thing should just be like a generic redirect, so we > don't need one installation per wiki etc... I basically think Q&A system within Wikimedia would be very useful. We can create one instance for whole Wikimedia (e.g. at help.wikimedia.org). The single site would have sections for different projects (Wikipedia, Commons etc.) within itself (kind of categorizing or tagging). Multilinguality could be solved within single installation too although I do not know if the above listed software supports it.
(In reply to comment #2) > Some software options: > - http://www.osqa.net/ (Python) > - http://www.question2answer.org/ (PHP 4.3+, MySQL 5) > - http://shapado.com/ (Ruby) > - http://www.lampcms.com/ > - http://pligg.com/ > - http://askbot.org Well I would be ruling at anything that doesn't run on PHP and mySQL since they most likely wouldn't be chosen for compatibility reasons. So that knocks out OSQA (Python), Shapdo (Ruby), LampCMS (MongoDB), Askbot (Python, I think)
(In reply to comment #5) > Well I would be ruling at anything that doesn't run on PHP and mySQL since they > most likely wouldn't be chosen for compatibility reasons. So that knocks out > OSQA (Python), Shapdo (Ruby), LampCMS (MongoDB), Askbot (Python, I think) You say that as you post to Bugzilla, which is written in Perl. ;-) This project/feature request is probably closer to OTRS (also in Perl) than a wiki. The big question is how much this would interact with current infrastructure (namely the current user auth system in MediaWiki). Would people want to use the same account? Would they settle for just being able to link their accounts (and what would prevent spoofing...)? Would accounts be completely unlinked? This all factors into what to choose. PHP is preferable from a maintenance standpoint (server architecture is built for a PHP app) and from a developer standpoint (plenty of PHP devs around).
I think the Teahouse project (http://meta.wikimedia.org/wiki/Teahouse) is aimed to tackle this issue.
Great idea! Something like question2answer or osqa would be a great addition / complement to the current Teahouse.
I think this would be a fantastic way to shift some of the non-developer support done for MediaWiki installations (like for end-users of third-party wikis, etc.) to a more direct place. I imagine we could recruit some folks who are otherwise not as engaged with WMF projects to help with some of the non-WM related assistance. In any case - I fully support the notion of a Q&A site for WMF projects and MW installations.
I don't think it should be a single one. "Why was my article deleted?" can't be answered without going down to the specific wikipedia instance. Not to mention the language barrier. I'd install it at http://en.wikipedia.org/help/ if any, but seems better served by an extension. Yet, I think a sidebar link leading to a LiquidThreads Help section is probably enough.
Kozuch: if you still have an interest in seeing this idea come to fruition, I'd recommend starting an RFC at mediawiki.org: <https://www.mediawiki.org/wiki/Requests_for_comment>. This would be somewhat similar to the page you've built at Meta-Wiki, but would much more heavily focus on the technical aspects of this bug. I think Platonides is probably correct in comment 10 that this bug is probably best served as a MediaWiki extension (written in PHP). This easily resolves a lot of the user auth issues mentioned in comment 6. On the other hand, an external tool/system would be easier to install and set up, but without any OAuth or LDAP or equivalent to link Wikimedia accounts to a separate auth system, it's gonna be rough to implement an external tool nicely (cf. OTRS, Bugzilla, Mingle, etc.). And I think if you're really interested in pursuing this idea, you should reach out to the [[Wikipedia:Editor engagement/Team]].
I think this system (or rather its part) should replace current http://en.wikipedia.org/wiki/Wikipedia:Help_desk.
Related: https://meta.wikimedia.org/w/index.php?diff=5055994&oldid=5041439&rcid=3835860 http://area51.stackexchange.com/proposals/49276/wikis
(In reply to comment #13) > Related: > https://meta.wikimedia.org/w/index. > php?diff=5055994&oldid=5041439&rcid=3835860 > http://area51.stackexchange.com/proposals/49276/wikis Thanks Nemo for those links. It would be good if Wikimedia expressed their support for the creation of this new Q&A website. Note that new editors can simply post their question on the site even if they do not have an account on the site. It is the "following" and "committing" part of the proposal that would require more experienced users with an account on the site to take part in, but all these for a simpler way for users to find answers to questions, the outcome is priceless.
(In reply to comment #14) > It would be good if Wikimedia expressed their support It's unclear to me who "Wikimedia" is. Do you mean Wikimedia Foundation? If so, see comment 11: An RFC might help.
CCing Steven and Quiddity. Is a Q&A system something that has been discussed? Is the lack of it a barrier for new editors? Setting priority as Lowest because this report has been silent for a year and nobody is currently planning to work on this.
(In reply to Quim Gil from comment #16) > Is a Q&A system something that has been > discussed? Only a few hundreds times. :-P If you mean within WMF, you can read Erik's https://lists.wikimedia.org/pipermail/foundation-l/2010-June/059052.html «With all this in mind, here are just a few concrete ideas for closing the gap: [...] Further experimentation with tools like IdeaTorrent for large-scale brainstorming and ranking purposes». In the meanwhile http://area51.stackexchange.com/proposals/49276/wikis was closed pending revival and sufficient support; and [[mw:Communication]] now includes a single address where all StackExchange related discussions are gathered. https://stackexchange.com/filters/116580/mediawiki
For what is worth, Phabricator has an application called Ponder that pretty much covers what is requested here: posting questions in relation to projects (topics) and allow people to reply and +1/-1 questions and answers. It would be trivial to enable it in our upcoming Wikimedia Phabricator if this is what we want to do. See an example at https://secure.phabricator.com/ponder/
... and now you can try it out at https://phab-01.wmflabs.org/ponder/
The Ponder on Phabricator might not be a bad solution. Better something than nothing :). But loads of traffic must be directed to such Q&A site so that some content is being created at all. I always have to laugh looking at the (mostly) useless discussions on mailing lists that have NO result mainly... a voting system on Q&A site will give you 1000 times better results in knowing what people want than tenths of mailing lists... lists really are history that belongs back to the 1990s...
(In reply to Jan Kucera (Kozuch) from comment #20) > a voting system on Q&A site will give you 1000 times better results > in knowing what people want Voting on followup comments to a question? Voting on questions themselves? I hope the former. (Personally wondering if Flow could have some up/down voting on non-initial comments in threads.)
(In reply to Quim Gil from comment #19) > ... and now you can try it out at https://phab-01.wmflabs.org/ponder/ Can a ponder item be transformed in a maniphest task?
(In reply to Andre Klapper from comment #21) > Voting on followup comments to a question? Voting on questions themselves? > I hope the former. (Personally wondering if Flow could have some up/down > voting on non-initial comments in threads.) Both. Ouestions tell you what needs to be solved (the problems) and answers give you solutions right away. Voting does the magic on both ends. So simple. Yet Wikimedia hasnt been able to acquire such simplicity since 2001 - its a very sad story. I mean the backstage now (=Meta) - consensus decisioning should still be used for content (Wikipedia articles). Look at Stack Exchage - that is what we need for whole Wikimedia.
If i18n needs to be part of the equation, then Flow with votes might be indeed more feasible than Phabricator with a multilingual environment. (In reply to Nemo from comment #22) > (In reply to Quim Gil from comment #19) > > ... and now you can try it out at https://phab-01.wmflabs.org/ponder/ > > Can a ponder item be transformed in a maniphest task? As of today there is no button for that. :)
To not only bite I add an idea - there probably should be multiple strictly THEMATIC Q&A site instances around - exactly as Stack Exchnge sites do it - i can imagime some for user help for a given language edition at a given project (like German Wikipedia) and then my favorite one for development (would complement Meta wiki). We lack the "development Q&A" very much since Strategy wiki was closed. I will start a RfC.
(In reply to Quim Gil from comment #18) > See an example at https://secure.phabricator.com/ponder/ If there is a Wikimedia-wide solution for this, I agree Phabricator (Ponder) would be good so there's not Yet Another Site. People will already be going there for issue-tracking at least (Maniphest).
