Jump to content

IRC office hours/Office hours 2012-12-15

From Meta, a Wikimedia project coordination wiki

[IRC Office Hours w/Sue Gardner, WMF Executive Director – https://meta.wikimedia.org/wiki/IRC_office_hours | Channel is logged http://bots.wmflabs.org/~wm-bot/logs/%23wikimedia-office/]
[11:29] <sgardner> What I don't know, and would be interested to know, is if the eventual default's intended to be VE mode as opposed to Read mode. I'm assuming yes, but I'm not sure.
[11:30] <James_F> Hey.
[11:30] <sgardner> hello James :-)
[11:31] <sgardner> Good timing :-)
[11:31] <James_F> Yes, but we're not absolutely sure it will work.
[11:31] <jps> it's a huge amount of javascript to push for everyone
[11:31] <James_F> But yes - long term, maybe we won't have a "read" mode at all.
[11:31] <James_F> jps: It's about 20kB extra on top of a 300kB payload, IIRC (and it'd be cached).
[11:31] <jps> I'm sure there will be a preference for logged in users
[11:31] <Amgine> <occasionally uses lynx on en.WP because it's so much faster>
[11:31] <sgardner> (This is where I asked: http://en.wikipedia.org/wiki/Wikipedia:VisualEditor/Feedback#VisualEditor_as_default)
[11:32] <Amgine> It's not the size, but the amount of time my netbooks take to process the js.
[11:32] <James_F> Amgine: Disabling JS is always something to consider (not just for Wikimedia sites).
[11:32] <jps> it's definitely worth measuring before leaping: http://www.extremetech.com/computing/110099-the-web-in-2011-html5-dominates-flash-trouble-for-data-capped-mobile-surfers
[11:33] <sgardner> So how do we want to handle office hours today? I'm assuming people want to talk about VE (so it's great that JamesF is here) -- do we want to also collect other topics?
[11:33] <James_F> jps: Indeed; when I said "might" this is exactly what I was talking about.
[11:33] <sgardner> (I am reading the usernames now to see who's here.)
[11:34] <jps> fundraiser, please. Specifically are we actually doing multivariate testing currently? If not, what do we need to do it?
[11:35] <sgardner> Raystorm, hello :-)
[11:35] <Raystorm_> hello sgardner :-)
[11:36] <sgardner> So probably VE, also the fundraiser ... what else?
[11:36] == Bence [bc060149@gateway/web/freenode/ip.188.6.1.73] has joined #wikimedia-office
[11:36] <sgardner> Bence :-)
[11:36] <Bence> hi :)
[11:37] <sgardner> Maybe we should just kick off, while people join and think about what they might want to talk about, with talking about the fundraiser?
[11:37] <Moonriddengirl> That sounds like a good idea, Sue.
[11:37] <sgardner> I am *so* happy with how it's going. I am really, really thrilled with Zack and the team. Has everyone got the basics, or do you want a recap?
[11:38] <Moonriddengirl> For the logs, a recap might be useful.
[11:39] <sgardner> (I am looking for a link to Zack's update on wikimedia-l, where he replied to Itzik. Moonriddengirl, do you think you could find that while I recap?)
[11:39] <Bence> btw, Zack wrote a summary on wikimedia-l yesterday, that is worth reading: http://lists.wikimedia.org/pipermail/wikimedia-l/2012-December/123043.html
[11:39] <Moonriddengirl> Sure.
[11:39] <Raystorm_> lol
[11:39] <sgardner> that's it -- thanks Bence :-)
[11:39] <Moonriddengirl> Yes, thanks, Bence. :D
[11:40] <jps> Zack asked for questions, so I asked the same one I've been asking since 2009: http://lists.wikimedia.org/pipermail/wikimedia-l/2012-December/123046.html basically, when will we ever use multivariate analysis to test all the volunteer-submitted messages from 2009
[11:40] <sgardner> So the gist of the fundraiser is that we annoy readers for money. So the goal is to annoy them as little as possible, while still generating the money the movement needs. The great thing about this year's campaign is that Zack and Megan have figured out how to annoy readers substantially less than ever before, while still increasing revenue. That's the basics of why this year is so good.
[11:40] <sgardner> Let me see if I can recap some numbers.
[11:41] <jps> historical numbers are at https://wikimediafoundation.org/wiki/Special:FundraiserStatistics
[11:42] <sgardner> Yeah, thanks jps. In 2010 the campaign ran 50 days, and in 2011 it was 46 days. This year we think it will be 16 days.
[11:43] <sgardner> Also: in 2010 and 2011 we ran in practically every country -- so, something like 200 countries. This year we are running in five English-speaking countries (I believe the UK, US, Canada, Australia and New Zealand), plus in France, Switzerland and Germany. So, only eight countries compared with 200 last year.
[11:43] <ori-l> at this rate we will reach singularity in a few years
[11:43] <sgardner> (I'll recap in a minute about how & when we'll fundraiser in additional countries, in April.)
[11:43] <sgardner> ori LOL.
[11:43] <Bence> what proportion of the total is to be raised in the non-8 countries?
[11:44] <Bence> also, what is the current goal? $35 mil?
[11:44] <sgardner> I hesitate to say, Bence, because I'm not sure I'll accurately remember. But I think Zack is planning to raise about 20 million in the winter campaign (from the five countries) and an additional three or five million in the non-8 countries. I might be misremembering though, so don't take that too seriously.
[11:45] <sgardner> (I am looking at the 2012-13 plan right now, to check my memory.)
[11:46] <sgardner> Yeah, we don't break out annual-campaign revenues from non-annual campaign revenues in the plan. But I think you're right -- it's roughly 35 million expected to come in via the campaign. Some of that I believe was already raised via testing prior to the campaign's official launch, and some will come in, of course, via the payment-processing chapters. I don't remember the specific target for the payment-processing chapters, off the top
[11:46] <sgardner> of my head.
[11:47] <Bence> ok, thanks
[11:47] <jps> about 300 volunteer-submitted fundraising messages from 2009 have not been tested, including about a dozen created by turning your goals statements into imperatives, sgardner: http://meta.wikimedia.org/wiki/Fundraising_2010/Messages/Mission#26_from_Sue.27s_goals_list_and_overlooked_esoterica
[11:47] <jps> we keep getting mixed messages about whether Megan H. has multivariate testing capabilities
[11:48] <sgardner> So we raised some money through testing, we will raise lots of money through the five English-speaking countries, we will raise some money via the payment-processing chapters in three countries, and we will raise the remainder (of the "many small dollars" money) via a second campaign in April. Let me talk about that for a second.
[11:48] <jps> can you please clarify whether or not she does, and if there is anything volunteers can do to help her get there if she doesn't?
[11:48] <sgardner> (yep, I will in a second, jps :-))
[11:48] <jps> thanks
[11:50] <sgardner> So about a month ago (?) Zack told me that the team was torn about priorities during the campaign. They could choose to 1) A/B test a bunch of messages in English, optimizing the banners and making them more effective in raising money with minimal annoyance to readers. Or, they could 2) focus on translation and localization for non-English languages, with the goal of optimizing those in the same way. It's a lot harder for the team to t
[11:50] <sgardner> est non-English languages, because everything needs to be translated before it can be tested. Testing in English is easy for them.
[11:51] <sgardner> They had had this split focus every year, and it is tough for them to manage. Rationally (if the purpose is purely to optimize the campaign) they would focus on English, because that's where most of the money comes from, but the risk then is that we do a bad job in non-English-language countries, and risk annoying readers in those countries more than we really need to, or want to.
[11:52] <sgardner> (Like for example, in English they are running the phrase "We don't take government money." We suspect that in most countries apart from the United States, that language doesn't resonate for donors, because in most countries people are less suspicious of government funding relative to Americans. But to know that for sure, it would need to be tested.)
[11:54] <sgardner> So upshot: Zack wanted to focus on optimizing for English during the winter campaign (because that's the most rational thing to do, including because it's easiest, and because that's where most of the money comes from, and because North America appears to be the geography that's most responsive to end-of-year giving) ... and then, once messages are generally optimized, later turn the team's attention to translating/localizing/customizi
[11:54] <sgardner> ng/testing in non-English languages.
[11:54] <sgardner> (Sorry this is long, I'm almost done.)
[11:54] <sgardner> And so that's what we're going to do. In April we will run a second campaign, in other countries -- I think focusing on Europe and Asia (?).
[11:55] <sgardner> And then after that, I am hoping, and Zack is hoping, that we can begin to fundraise flatly all year (rather than focusing on a single winter "campaign") -- with continual testing and iterating, and with readers seeing a few banners once in a while, rather than a concentrated push of banners once a year.
[11:55] <Risker> I for one think that splitting the fundraising campaigns this way is an excellent idea
[11:56] <jps> rationally, it would seem to me that you would test in English first, take the top-N, translate them, and see if they do as well as known controls from the past in other languages
[11:56] <sgardner> Fundraising "flatly" would spread the work throughout the year, making it easier for Zack's team, and would also reduce the risk that one year a terrible natural disaster or some such might derail our campaign. (That's not a likely possibility, but it could happen and we're conscious of it.)
[11:57] <sgardner> Yes jps, I think that's what they're planning. So they take some time now to test and determine what works best, then translate and optimize those top N.
[11:57] <sgardner> Oh one other thing! This year we are even less dependent on Jimmy than ever before.
[11:57] <Bence> did Jimmy write a new appeal this year, or are you using the one from the previous year?
[11:58] <Amgine> Actually, international marketing suggest that model will have a low percentage of success as nearly as many American themese are antithetical as are successful in other venues.
[11:58] <sgardner> In 2010 we ran 36 "Jimmy days." In 2011 it was 12. This year I am not sure how much Jimmy we've run, but it's practically none. I think some of the chapters may use the Jimmy messaging still sometimes, but I don't think the WMF is using it at all.
[11:58] <jps> the idea that plain text messages do better than Jimmy's appeal is additional evidence to test the remaining volunteer-submitted proposals
[11:58] <sgardner> Amgine we'll see. There are definitely differences between what resonates for Americans vs the rest of the world. Like what I said about "government," above.
[11:59] <Ironholds> indeed
[11:59] <Ironholds> I've had people from europe email in
[11:59] <Ironholds> "why won't you take government support?"
[11:59] <Ironholds> "you could do so much!"
[11:59] <sgardner> jps: we think what is working is facts, rather than appeals to emotion. We also think there's real value in getting facts into the banners (rather than requiring a click-through) because that way more people are exposed to the basic messaging -- we're a non-profit, etc.
[11:59] <Ironholds> because government is less mistrusted in a lot of places.
[12:00] <sgardner> yes exactly Ironholds. The first time I saw the "no govt money" messaging I was like "I'm Canadian. I trust my government" :-)
[12:00] <sgardner> So that's the gist. I think I'm done. I can answer jps's question now, and then maybe other people have things about the campaign they'd like to talk about?
[12:00] <Ironholds> sgardner: you shouldn't when the governor-general is involved ;p
[12:00] * Risker trouts Ironholds
[12:01] <JohnLewis> Some government you can trust. Others you cant. The last government (For Briain), You cant trust them with a penny. Unless you want a world war.
[12:01] <Ironholds> Risker: you're just impressed I know the recent history of Canada in government terms.
[12:01] <Risker> well, I think the American perception of government funding is somewhat equated with government control
[12:01] <Risker> which is something of an issue
[12:01] <Ironholds> ya
[12:02] <Amgine> Uhm, this is straying from the topic of fundraiser.
[12:02] <jps> do you feel comfortable drawing such inferences (i.e. facts do better than appeals) given the large variance between messages, and the fact that appeals have traditionally performed best?
[12:02] <sgardner> :-)
[12:02] <jps> wouldn't measuring them all at once with a large multivariate test be much less risky?
[12:03] <sgardner> jps: yeah, I do. Appeals performed best until we started testing the facts banners. The facts banners do dramatically better. (I don't have the numbers, but they're on meta.) And we have such large numbers of people seeing the banners, we can be confident about the test results.
[12:03] <Amgine> SGardner: would it be useful to organize 'campaigns' to allow continuous meta-analysis of performance? Also, does our ad server software manage campaigns?
[12:03] <Ironholds> jps: on multivariate analysis: I suspect that the analytics used on the fundraiser is probably not something sgardner can speak to directly :)
[12:03] <sgardner> (I wrote a test banner with Zack the other night that performed 13% better than the control. That was pretty pleasing.)
[12:03] <sgardner> Ironholds, that is true :-)
[12:03] <Ironholds> unless she's secretly been dressing up as Katie two days a week
[12:03] <Ironholds> which would be /weird/
[12:04] <sgardner> I would totally do that.
[12:04] <sgardner> I would like to have Katie's hair.
[12:04] <Ironholds> sgardner: now you know your halloween outfit for next year.
[12:04] <sgardner> Sorry, OT :-)
[12:04] <Ironholds> she does have great hair. Anyway.
[12:04] <ori-l> sgardner: me too
[12:04] <sgardner> :-)
[12:04] <Ironholds> okay, next halloween
[12:04] <Ironholds> everyone comes in as Katie.
[12:04] <Ironholds> everyone.
[12:04] <Risker> meanwhile.....
[12:04] <sgardner> Oh, Amgine, I'm not sure I can answer that question. Can anyone here help with that?
[12:04] <Risker> Visual Editor?
[12:04] <jps> well, it's frustrating to be told every year that the volunteer submissions will be tested, but they haven't been yet. The variance is quite large. There are untested gems waiting
[12:05] <Moonriddengirl> jps, I see you asked Zack a question about this about two hours ago on wikimedia-l. I hope he'll be able to answer that question.
[12:05] <jps> whenever I ask Zack or anyone else on the fundraising team, they say that they will test them. How long should I wait before giving up?
[12:06] <ori-l> jps / Amgine: I can't say much about fundraising's approach to testing, but my experience has been that that people often chronically underestimate just how hard it is to run robust experiments
[12:06] <sgardner> Last year Zack wanted to run a so-called "contribution campaign" or "participation campaign" in addition to the campaign for donations. I think that soon he will finally be able to do that. And I think that's where volunteer "appeals" might be more useful than in appealing for donations. Basically, now that we are going to (finally!) have the VE, that will reduce the single most immediate barrier to editing (wiki syntax) and thereby ma
[12:06] <sgardner> ke it possible for more people to edit. So, a contribution campaign using the sitenotice is something we are thinking about doing.
[12:07] <jps> well we know that there are multivariate tools available. An 8-way test was run in May 2012. October had two three-way tests. I've examined the javascript harness for gathering stats. What concerns me is that apparently Megan doesn't even have a file with all the tested and untested phrases in it
[12:08] <sgardner> (I would really like to see that happen. The sitenotice is an incredibly powerful tool, and we have a testing infrastructure which would work just as well for inviting people to edit, as to donate.)
[12:08] <jps> can it really be true that nobody has bothered to compile the message phrases into one file with scores for the previously tested phrases?
[12:09] <Moonriddengirl> Sue, we're over half time now. Maybe we should talk a bit about the visual editor to be sure we don't run out of time? It's a pretty massive development. :)
[12:09] <sgardner> Moonriddengirl: okay!
[12:09] <sgardner> Is JamesF still here?
[12:09] <Ironholds> James_F: ping ;p
[12:09] <sgardner> :-)
[12:09] <James_F> I am, sorry.
[12:09] <sgardner> So do people want to talk a little about their VE experience so far? Are people using it, and reporting bugs?
[12:10] <sgardner> (John Broughton has been *great*, I've noticed.)
[12:10] <Ironholds> and Raymond, too.
[12:10] <James_F> Lots of people have given very helpful feedback; thank you to all of them. :-)
[12:10] <Ironholds> I gave him a barnstar from my official account for not only feedback-ing but helping triage it into bugzilla. He wins points :)
[12:11] <sgardner> I was surprised to see you triaging or coordinating feedback, Ironholds. How did you get roped into that? :-)
[12:11] == Fluffernutter [~Fluffernu@wikipedia/Fluffernutter] has joined #wikimedia-office
[12:11] <sgardner> Fluffernutter! :-)
[12:11] <James_F> sgardner: I twisted his arm. :-)
[12:11] <Fluffernutter> hi :)
[12:11] <sgardner> I knew it, JamesF :-)
[12:12] <Risker> James_F, I don't think I've been so genuinely excited by a MediaWiki software development in all the years I've been around
[12:12] <James_F> Risker: Thank you. I think? :-)
[12:12] <James_F> Risker: There's still a long way to go, however.
[12:12] <sgardner> It's absolutely fabulous. It is my favourite thing :-)
[12:13] <Risker> let's put it this way. The second edit I did had a bug, and I was still thrilled, James_F
[12:13] <sgardner> JamesF, do you want to talk a little about your thinking about further roll-out?
[12:13] <James_F> Risker: A bug we fixed! Though not deployed yet.
[12:13] <James_F> sgardner: Sure.
[12:13] <Ironholds> sgardner: I'm the only community person in Engineering as a whole
[12:13] <Ironholds> I get roped into a lot of stuff ;p
[12:14] <James_F> So, two deployment items: deploying to users more directly (not opt-in but 'actually' deployed for all users, in a sliding scale upto being on all the time, possibly) and deployment to other wikis than the English Wikipedia.
[12:14] <Raystorm_> yeeeeees, other wikis! please :D
[12:15] <Fluffernutter> (oh, since I'm late I don't know if this has been mentioned yet, but I want to compliment James_F and team SO MUCH for giving up an opt-in here rather than an opt-out)
[12:15] <James_F> For the second of these, there are three things holding us back from deploying to more wikis - first, we need to get Parsoid (the service VisualEditor relies on) to a place where it can handle non-English content.
[12:15] <James_F> Second, we'd need to get VisualEditor checked in these languages, and right now we already know of a few bugs that are big enough that it wouldn't be hugely respectful to the communities to dump it on them.
[12:16] <James_F> Third, well, there are lots of specialisations in non-Wikipedia wikis, and there's a huge amount of work to do in getting VE working for those.
[12:16] <James_F> Be that source-alongside-content for Wikisource or galleries for Commons or ... lots of things.
[12:17] <James_F> For the first deployment issue, right now we're looking mostly for feedback from people about what works and doesn't, what they see working in their workflow or being an issue, and getting them familiar.
[12:17] <James_F> Moving to the next phase, where we might reasonably expect people to use it on a day-to-day basis for editing, will take some time.
[12:17] <Amgine> Have you considered en.wiktionary, which has extensive non-english content?
[12:18] <Ironholds> Amgine: that's actually a /really/ good idea
[12:18] <Ironholds> and the burden on us to translate feedback would be much reduced.
[12:18] <James_F> Once we're there, we will probably want to move to having the VisualEditor switched on for all people, accounts and non-accounts.
[12:18] * Ironholds strokes beard, remembers he shaved, curses.
[12:18] <James_F> Amgine: Considered and dismissed.
[12:18] <gwicke> afaik non-English content is not much of a problem
[12:19] <James_F> Amgine: Right now we don't support template editing, which means for Wiktionaries it's essentially useless.
[12:19] <gwicke> it's non-English namespaces, link trail regexps, language variants etc
[12:19] <Amgine> <grins> touché
[12:19] <James_F> Yeah, gwicke knows more about the specifics than me (and is one of the heroes that built it). :-)
[12:20] == lyzzy [~lyzzy@wikimedia/lyzzy] has joined #wikimedia-office
[12:20] <James_F> And finally, at some point (quite a way off), we will probably have a "read" tab with the VisualEditor in it and an "Edit Source" tab that takes you to the existing wikitext edit.
[12:20] <James_F> Err. Editor.
[12:20] <James_F> Anyway.
[12:20] <James_F> Any questions? I know I've been just spamming for a few minutes. :-)
[12:20] * Ironholds raises hand
[12:21] <sgardner> (You might talk about timelines a little James_F, to the extent that you can. Once other people get stuff answered.)
[12:21] <Ironholds> does wikidata, in your view/in theory, solve for any problems with things like templates?
[12:21] * James_F nods.
[12:21] <Moonriddengirl> (Office hours is over in about 10 minutes. If you have any final questions, you might want to go for it.)
[12:21] <Ironholds> I appreciate the answer is probably "no". But I thought I'd ask :)
[12:21] <James_F> Ironholds: For VE? No. For editing communities, yes. VE will integrate with Wikidata (like with other 'special' content), but it won't necessarily make things much easier.
[12:22] <Ironholds> *nods*
[12:22] <James_F> Ironholds: It will mean that templates when looked at will still have parameters, it's just that those parameters will be set as Wikidata query calls instead.
[12:22] <Ironholds> cool
[12:22] <James_F> sgardner: Sure.
[12:22] <sgardner> (I am kind of concerned about our ability to balance "community happiness with VE" against "getting VE rolled out as the default.")
[12:22] <jps> I think that templates are too hard to expect VE to solve. Parsing them is already equivalent to solving the halting problem. Therefore they should simply be displayed in raw wikitext.
[12:22] <Ironholds> sgardner: I am actually working on this problem :)
[12:22] <James_F> jps: I think you're overly-pessimistic.
[12:22] <Ironholds> complete systems analysis of the edit workflows looking at what human elements break and burn out if traffic goes up.
[12:23] <gwicke> Ironholds: there is not much danger of wikidata being used on popular wikis without a solution to the caching problem
[12:23] <jps> James_F: I am certain that I can prove that parsing them is equivalent to the halting problem
[12:23] <James_F> Well, right now we're still in the "research" phase, and there are a bunch of things we'll need to do that we're not sure how it will work yet.
[12:23] <James_F> jps: I am certain that you are right, but are thinking of the wrong problem. :-)
[12:23] <Amgine> gwicke: Is there a parser specification?
[12:23] <sgardner> (Ah, very good Ironholds. Because there's going to come a point --which we are nowhere near, yet-- where it wants to be rolled out as the default, despite being imperfect. Because its benefits will outweigh any small problems or brokenness.)
[12:23] <gwicke> Amgine: depends on what you mean with that
[12:24] <gwicke> Amgine: the DOM is specced in http://www.mediawiki.org/wiki/Parsoid/MediaWiki_DOM_spec
[12:24] <jps> I mean, parsing them into any reasonable set of attribute-value pairs. I'm not talking about theoretically. There are already templates that look at parameter values to decide between {{{1|}}} and a named parameter, for example
[12:24] <aude> what, office hours?
[12:24] <Ironholds> sgardner: agreed. But we won't know how much work we need to do to fix the impact until we know how big the impact is likely to be
[12:24] <aude> hi
[12:24] <James_F> So timelines are hard to pin down. Right now we hope to get to a better state with some early "inspectors" - references, categories, images, etc. - in the next few months.
[12:24] <Ironholds> so I decided to start the project now to avoid "we know what'll happen and don't have time to solve for it" or whatever.
[12:24] <Moonriddengirl> hi, aude. :) We're wrapping in about 5 minutes. :D
[12:25] <gwicke> jps: that part is pretty much solved
[12:25] <James_F> We seem to be in the business of quarterly milestones, so roughly March-ish for that.
[12:25] <jps> what a relief?
[12:25] <James_F> But I don't want to raise false expectations and promise grandiose targets that we'll miss.
[12:26] <James_F> Anyway. Any more questions?
[12:26] <James_F> (Or re-state a question if I've skipped/missed it!)
[12:26] <jps> there's also complexity such as what happens when someone changes a template while the VE has an instance open for editing. Why not just let templates fall back to wikitext first, roll that out for the vast majority, and then work on perfection from that good?
[12:26] <Ironholds> sgardner: I'd argue that there's going to be one /big/ problem, which is what happens to all the community workflows when we triple or quadruple the entry numbers.
[12:26] <Ironholds> it's that I'm concerned about more than technical errors.
[12:26] <James_F> jps: Have you used the VisualEditor yet? That's where we are right now. :-)
[12:27] <aude> reminder we have wikidata office hours on tuesday at 16:30 UTC :)
[12:27] <jps> I used it yesterday and it wouldn't let me edit templates. They had diagonal bars. How do I get their wikitext within VE?
[12:27] <sgardner> Ironholds: yes. That's an unknown at this point -- we have no idea how much of an influx we're going to create. But yes, we are probably all a little anxious about opening the floodgates :-/
[12:28] <James_F> jps: You don't; that will come, hopefully soon (though there's some seriously-complicated engineering to match changed wikitext to points in a translated HTML DOM, but that's not a useful topic for this slot).
[12:28] <jps> cool
[12:29] <Ironholds> sgardner: Indeed. I'm running different models - 100 percent, 200 percent, 500 percent increase, so on.
[12:29] <aude> sgardner: stuff like the status updates on mediawiki.org see more experiments by anons
[12:29] <aude> those have an editing widget
[12:29] <aude> i expect more experiments with VE
[12:30] <aude> https://www.mediawiki.org/wiki/Git/Conversion <-- here for example
[12:30] <sgardner> So, me recapping & making up some stuff to see if it's true, James_F....... Currently people are testing & we're collecting bugs and other experiences, hopes, etc. Within a month or so (?) we'll know roughly the size and scope and scale of brokenness, and roughly what needs to be fixed before VE could be default for English, and we'll also be beginning to roll it out as alpha for some other languages including RTL, so we can start know
[12:30] <sgardner> ing the size-and-scale for them. Is that roughly right, or am I just making up stuff here?
[12:31] <James_F> sgardner: The first bit is right; we can make the second half right, but that will mean changing priorities in Parsoid - gwicke can advise, but we could do that if needed.
[12:31] <James_F> sgardner: But more important to fix the big issues first before inflicting them on RTL wikis when we already know it's broken, I feel.
[12:32] <sgardner> Nah, I'm not trying to derail the planning. You should stick with your plan. I'm just testing my assumptions/guesses.
[12:32] <gwicke> James_F: we suppot RTL already fine- we actually don't care about it
[12:32] <Amgine> script-agnosticism is good.
[12:32] <James_F> gwicke: Sorry, yes, I meant "other wikis like those with RTL".
[12:32] <gwicke> it's wikis with different language defaults that would need to be supported
[12:32] <James_F> Yes.
[12:33] <gwicke> that's pretty easy to do, but right now our priority is making editing on English wikis more solid
[12:33] <sgardner> So I know we need to wrap up soon. So I want to just take a second to say how thrilled I am about the VE. It is such an enormous step, and it's making me really happy :-) Happy it's launched as alpha, happy it's being tested, happy to see a plan develop for the next six months :-)
[12:34] <gwicke> might get to it in two weeks or so
[12:34] <Moonriddengirl> It's a pretty amazing development. :)
[12:34] <Moonriddengirl> Any final quick questions or comments, anyone?
[12:34] <James_F> Thank you (mostly, thank you to my team who've done the great work). :-)
[12:34] <Bence> I'd just like to spam the end of the log with AffCom's recruitment call: http://meta.wikimedia.org/wiki/Affiliations_Committee/Call_for_Candidates_2013
[12:35] <sgardner> ah, good idea, Bence :-)
[12:35] <Moonriddengirl> Okay, then. It looks like we're a wrap. Thank you, Sue, and thank you everyone for taking part of your Saturday. :)
[12:35] <aude> thanks sgardner
[12:35] <aude> happy weekend :)
[12:35] <Moonriddengirl> You, too, aude!
[12:36] <sgardner> Thanks everybody. I like doing these on Saturday -- it works for the largest number of people internationally, and it's actually easier for me than fitting it into weekdays. Thanks to the WMF people who turned up during their weekends, too :-)