IRC office hours/Office hours 2014-06-19

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

[15:00:11] <Elitre> looks like we're getting started!
[15:00:34] * James_F waves.
[15:00:37] <Elitre> Hello everybody, and welcome to the June office hour for VisualEditor.
[15:01:13] <Elitre> James_F: welcome! as usual, VE's product manager is our guest star.
[15:01:33] <Guerillero> maggie claims I can't be too mean to the staff today
[15:01:34] <Guerillero> :P
[15:01:49] <Maggie_Dennis> :D
[15:01:55] <Elitre> I'm Erica - Elitre (WMF), a community liaison, and I'll be facilitating this office hour today.
[15:01:55] <jayvdb> it has been a year since launch ?
[15:01:58] <James_F> Hey everyone. Happy to take your questions, thoughts, complaints, concerns, ideas, likes, etc. :-)
[15:02:17] <Elitre> jayvdb: welcome! and yes, more or less.
[15:02:21] <Maggie_Dennis> Thank you for your kindness, Guerillero. :)
[15:02:29] <Guerillero> your welcome
[15:02:51] <jayvdb> what is the current percentage of edits using VE ?
[15:03:11] <Elitre> So, who has questions? Jayvdb is first, it seems.
[15:03:29] <James_F> jayvdb: Hey. Yes, it's just under a year since the initial "on by default" switch for a big set of wikis.
[15:03:51] <jayvdb> and since we have two wonderful community liaisons facilitating, a question for them: has the rate of bugs / problems / gaps in user expectations started slowing down ?
[15:03:57] <James_F> jayvdb: It's been just over two years since VisualEditor went live with original community testing (initially on MediaWiki.org).
[15:04:30] <James_F> jayvdb: The usage of VisualEditor is mixed between different wikis, and different user groups, but to give a brief summary:
[15:04:37] <Maggie_Dennis> jayvdb, I'm afraid that I'm not active in that area; I'm just here. :D But I'm sure that Erica and James can reply to that one.
[15:05:06] <James_F> On wikis where VisualEditor is on by default, anonymous users use VisualEditor for ~ a third of their content edits.
[15:05:29] <jayvdb> Ah Maggie_Dennis , .. that is their loss ;-)
[15:05:33] <James_F> Users of registered accounts created after the switch to default use it for ~ a fifth.
[15:05:38] <Maggie_Dennis> You charmer. :)
[15:05:53] <James_F> Users of registered accounts created before the switch to default ("old hands" or whatever) use it for ~ a twentieth.
[15:06:32] <James_F> So, as you can see, adoption isn't as high as we'd want and we're looking to make VisualEditor better so that more people choose to use it over wikitext.
[15:06:38] <jayvdb> Im not often called a charmer :P
[15:06:52] <James_F> (The alternative of course would be to make wikitext worse, but that wouldn't be very ethical. ;-))
[15:07:05] <jayvdb> James_F: do you have any targets for adoption ?
[15:07:29] <James_F> However, slightly more anonymous users use VisualEditor in any given day than those who use wikitext.
[15:07:36] <Elitre> jayvdb: third question noted :)
[15:08:37] <James_F> So to speculate, the old-hand IP editors (who we know exist but can't label in our data) look like they're more likely to use wikitext, and more likely to be productive in the number of edits.
[15:08:44] <James_F> But we don't know for sure.
[15:09:48] <James_F> In terms of targets (to answer question three before Elitre can do question two?), ish. "Higher", certainly. I think we can probably convert IPs over to a 2:1 ratio, at least.
[15:10:00] <jayvdb> James_F: your responses are a bit confusing. first, do you have an overall percentage of edits on all wikis? secondly, "anonymous users use VisualEditor for ~ a third of their content edits" vs "slightly more anonymous users use VisualEditor in any given day than those who use wikitext" dont line up - could you explain how these two statements are both accurate.
[15:11:16] <russavia> what's with all the questions jayvdb, you're like the kid in uncle buck
[15:11:21] <Elitre> I think James just answered your question, jayvdb? kind of edit conflict there :)
[15:11:49] <James_F> jayvdb: On all wikis? I could ask Analytics to crunch that number, but I don't have it to hand. On wikis where VisualEditor is active, I believe that VE is ~ 6.5% of all edits (but that counts talk page edits, bot edits, etc.) so it's not really a valuable number to talk about. Post-Flow maybe we can have a better idea, but…
[15:12:14] <Elitre> thanks for the "wonderful", jayvdb :D I wanted to answer to your previous question: I feel like we're almost not talking about the same product here. I wish people would give it another try if they haven't had a chance to do that lately.
[15:12:35] * James_F nods.
[15:13:15] <James_F> To jump in with my perspective, the rate of community-filed bugs has certainly dropped, as has the severity (in general).
[15:13:54] <James_F> I can't really speak to people having "gaps in user expectations […] slowing down" but certainly people are in a more positive frame of mind after using VisualEditor now than they were before.
[15:14:09] <James_F> jayvdb: Not sure if that answers your question(s)?
[15:14:20] <Elitre> I have recently seen a *tracking* bug being closed; I was impressed.
[15:15:22] <James_F> Yeah, we're relatively aggressive in fixing bugs we find, so often people won't notice the issue before it's already solved.
[15:15:28] <James_F> Just a moment, I had some numbers
[15:15:50] <Elitre> we ain't going anywhere, James_F...
[15:16:28] <Krenair> Elitre, https://bugzilla.wikimedia.org/show_bug.cgi?id=56865 ?
[15:16:42] <Revi> any updates on bug 50631?
[15:16:49] <Revi> @link bugzilla:50631
[15:16:49] <wm-bot> https://bugzilla.wikimedia.org/show_bug.cgi?id=50631
[15:17:01] <James_F> Aha, yes: in the last quarter we made about 800 individual changes to the code (some of which are smaller than others), fixing about 300 valid bugs and closing another 150 invalid ones (duplicates etc.); in that time, we had about 380 bugs filed, mostly by the team ourselves.
[15:17:15] <James_F> So in net terms the number of open bugs is going down.
[15:17:17] <Elitre> Krenair, nope! I think the one I saw had to do with media.
[15:17:26] <Elitre> Revi: question noted.
[15:17:47] <Revi> ty
[15:18:13] <Elitre> Krenair: you are not going to open any of them just to make me sad, are you? :)
[15:18:17] <Krenair> Elitre, well that makes two :) Unfortunately there is a strange problem with one of the descendent bug's patches, it had to be reverted
[15:18:20] <Elitre> *re-open
[15:18:23] <James_F> That's not a totally fair picture, because we don't always create bugs for new features until just before we fix them.
[15:18:55] <James_F> If I just sat down and started writing out crazy ideas, we could probably add another 1000 enhancements to the backlog, but it might not be so valuable. :-)
[15:19:00] <Krenair> Elitre, https://bugzilla.wikimedia.org/show_bug.cgi?id=50341 's grandparent is that tracker I mentioned
[15:19:19] <James_F> Revi: Hey there.
[15:19:38] <Revi> yes?
[15:19:51] <jayvdb> James_F: I would appreciate it if you could provide that statistic in future office hours. I will ask it again ;-)
[15:20:09] <James_F> Revi: That bug is a really horrible one, yes; it's caused by the different ways that users' operating systems insert characters in Korean (and, more generally other languages using IME).
[15:20:36] <Revi> hmm.
[15:20:38] <jayvdb> Elitre: I'll give it another round of testing this month. thx
[15:21:06] <James_F> Revi: We've been working on (several) complete re-writes of character inserting since then, many of which we've landed since then.
[15:21:30] <James_F> Revi: But code fixes and bug fixes are different, sadly; I believe that that one still has some edge cases where it doesn't work.
[15:22:01] <Elitre> jayvdb: I'm pretty sure you know this already, but for others who don't: there's a bunch of dashboards at https://meta.wikimedia.org/wiki/Research:VisualEditor .
[15:22:16] <James_F> Revi: The Language Engineering team's David Chan is working on trying to get IME support working fully in VisualEditor, even for the more unusual cases like iBus's Korean inserting.
[15:22:42] <Elitre> jayvdb: great! please let us know how it goes?
[15:23:01] <James_F> Revi: Unfortunately that just means "it doesn't work yet", sorry. :-( We think we may be near to a break-through that can work for all IMEs, but it's going to need a huge amount of testing to confirm.
[15:23:01] <Revi> Good to know
[15:23:16] <Revi> hmmmmm ok.
[15:23:26] <James_F> Revi: It's really critical that users of any language can use VisualEditor to write what they want, with the tools that work for them.
[15:23:53] <jayvdb> Elitre: on that dashboard, I cant find a way to view % of all edits on all wikis
[15:24:39] <James_F> jayvdb: As I said, that number doesn't exist.
[15:25:07] <James_F> jayvdb: We could ask Analytics to calculate it instead of other work they're doing, if you think it's really critical. I don't.
[15:25:18] <Elitre> Krenair: looks like I commented on that https://bugzilla.wikimedia.org/show_bug.cgi?id=50341 back then. Surprise, shock, etc. :p
[15:25:30] <jayvdb> with or without che caveat "on wikis where VisualEditor is active" which is IMO a cop-out as it is bugs preventing further rollout, so % of edits being a reasonable measure of effectiveness of this project, wikis it cant be enabled on are non-effectiveness
[15:25:34] <mvolz> James_F: an alternate theory for your anonymous users; users using VE make longer edits, whereas wikitext users break-up these edits into several separate edits. Do you we have stats on the length of each edit?
[15:25:52] <russavia> i would have thought that such statistics would have been part of the entire process
[15:26:11] <James_F> jayvdb: For example, no Wikidata edits are made with VisualEditor (and can't be); skewing the data with millions of spurious entries is misleading at best.
[15:27:08] <jayvdb> ok, then the caveat should be: in all namespaces where the content model is wikitext
[15:27:45] <James_F> jayvdb: Filtering out bot edits, bot-like edits, talk-page edits, LiquidThread edits, Wikidata edits and hundreds of other non-VE-possible edits is probably not a trivial matter to ask Analytics to spend time on instead of more important questions like "what is the effect of new data centre our reading speeds for uses in Asia" or similar.
[15:28:28] <James_F> jayvdb: You'd also need to filter out the wikis which don't have VisualEditor enabled (which changes from week to week, so the data would need to be multi-filtered over time).
[15:28:41] <James_F> jayvdb: Can the number be calculated? Yes. It is worth the effort? I doubt it.
[15:28:56] <jayvdb> which wikis have VE disabled on a week to week basis ?
[15:29:27] <James_F> jayvdb: The list of which wikis have VE enabled changes on a week to week basis, e.g. Meta got it a few weeks ago.
[15:29:51] <James_F> jayvdb: So we'd need to discount meta edits until VisualEditor was enabled there, but then count it.
[15:30:02] <russavia> james_F by ^ ^ do you mean there are projects that have VE enabled one week, and then not the next?
[15:30:02] <James_F> jayvdb: Does that make sense?
[15:30:32] <Elitre> https://www.mediawiki.org/wiki/VisualEditor/Rollouts is the list of wikis which already have VE, and those waiting patiently to get it.
[15:30:57] <James_F> russavia: No. There are two wikis where it was enabled for all users and is not any more (English Wikipedia in September and Spanish Wikipedia in January).
[15:31:13] <russavia> ok thanks for clarifying that
[15:31:51] <jayvdb> James_F: sorry but I agree with russavia - the investment in developing an appropriate dataset to ask these types of questions should be part of 'measurement & reporting' of *this* project
[15:32:02] <James_F> mvolz: That's a plausible reason, yes (and the related "VisualEditor means not needing to do follow-up fixing edits"). We don't have edit length data, no.
[15:32:09] <Elitre> (not that all of them are so patient. we get requests to please please please bring it to a certain wiki, from time to time. it's exciting.)
[15:32:56] <James_F> Elitre: Yeah; my apologies to those wikis that we've kept waiting whilst we make certain aspects better for them. :-(
[15:32:59] <jayvdb> that you might outsource it to the Analytics team is a matter of details. Given the cost of this project, an overall effectiveness measure is needed. The goal is to replace at least all source edits in content namespace
[15:33:52] <russavia> i've used VE a few times -- i don't like it -- how is such usage amongst editors like that gauged? surely having a percentage of edits would be a very effective way of gauging the real uptake by the wider editorial community
[15:34:00] <James_F> jayvdb: Thank you for your suggestion about how Analytics should prioritise their time; I will make sure it is passed on to them.
[15:34:24] <russavia> james_f -- ^ ^ no need for sarcasm
[15:34:25] <jayvdb> It has always been enabled on English and Spanish - it is just not the default.
[15:34:34] <James_F> russavia: You may have missed my earlier answers – the dashboards for each wiki give the percentage edit rate for that wiki.
[15:34:51] <James_F> russavia: What sarcasm? Sorry, I'm lost. Could you explain?
[15:34:59] <jayvdb> James_F: I didnt see an answer to > secondly, "anonymous users use VisualEditor for ~ a third of their content edits" vs "slightly more anonymous users use VisualEditor in any given day than those who use wikitext" dont line up - could you explain how these two statements are both accurate.
[15:35:19] <James_F> jayvdb: Sure, I'll repeat my comment explaining that:
[15:35:47] <James_F> "<James_F> So to speculate, the old-hand IP editors (who we know exist but can't label in our data) look like they're more likely to use wikitext, and more likely to be productive in the number of edits."
[15:36:03] <James_F> Well, "explaining" is too strong.
[15:36:11] <James_F> Speculating. We don't know why for sure.
[15:36:48] <mvolz> jayvdb: so, each individual anonymous wikitext editor is making lots of edits
[15:37:00] <mvolz> and each individual anonymous VE editor is making fewer edits
[15:37:07] <Elitre> Russavia, I think the point is that it's not up to James to make that decision - analytics has its own director.  I don't believe he intended that to be sarcastic.Â
[15:37:12] <James_F> jayvdb: And mvolz's alternative explanation is just as good (perhaps better).
[15:37:14] <jayvdb> ah, thanks for clarifying that mvolz
[15:37:26] <mvolz> and there are a slightly larger number of VE users than wikitext editors
[15:37:34] <Elitre> (not that he can't - he's British! But I think I could tell if he was.)
[15:37:36] <Elitre>  :)
[15:37:37] <russavia> ok, so if that's the case i don't see why it should be so hard to do what jayvdb is suggesting -- its just creating an extra subset of data, and it would surely be beneficial to someone -- that jayvdb thinks it would be useful, i'm inclined to believe that if he sees a reason for it, then there probably is -- and isn't simply a waste of time -- so please do pass the suggestion along :)
[15:37:53] <jayvdb> Looking at http://ee-dashboard.wmflabs.org/graphs/enwiki_ve_hourly_perc_by_ui , it appears that 0.5% is the rough percentage of VE edits on English Wikipedia
[15:37:55] <James_F> Yeah. One of the things we've seen a lot with wikitext editing is people making edits find they broke something, so immediately re-edit the page to try to fix it (often without success the first time).
[15:37:57] <russavia> elitre: thanks
[15:38:53] <James_F> jayvdb: Indeed, opt-in features without much current fanfare get little usage.
[15:39:16] <jayvdb> James_F: do you know which wiki has the highest percentage of VE edits?
[15:40:03] <James_F> jayvdb: One of the things that would be really helpful would be if we could get more real-world usage of VisualEditor on the English Wikipedia so we make sure we don't make changes that go against how that wiki would like things.
[15:40:55] <James_F> jayvdb: I believe Polish hovers around 20%; French and Italian around 15%.
[15:41:07] <Elitre> VE being opt-in on some wikis also means most people do not realize at all when a bug they cared about has been fixed, or when a new, ground-breaking feature is out there. :/
[15:41:11] <James_F> jayvdb: Which, as I said, is not anything like as high as we'd want.
[15:42:06] <James_F> Elitre: Good point. Feedback is needed to make things better, otherwise we might be chasing our tail.
[15:42:46] <jayvdb> James_F: have you any suggestions on how to get more real-world usage on en.wp? e.g. maybe enable it by default in the draft namespace ? do you have other gradual deployment options in mind?
[15:43:23] <James_F> jayvdb: I think enabling it by default on a specific namespace (and not the others) might be a bit jarring for users, but we could do that.
[15:43:45] <russavia> hasn't it been a case of chasing ones tail since the beginning though? seriously speaking
[15:43:54] <mvolz> James_F: we might not have length data, but we have the time series, correct? So if we're seeing successive corrective edits, you might be able to tease that out statistically and correct for it. Someone tell analytics! :)
[15:44:51] <jayvdb> im not seeing 20% on http://ee-dashboard.wmflabs.org/graphs/plwiki_ve_hourly_perc
[15:45:03] <James_F> jayvdb: An alternative is to enable it by default for 10% of all anonymous users (picked randomly), and shift that to 25%, then 50%, then 100% – but doing so for anonymous users in advance of logged-in users might be hard for the community (lots of anons asking editing questions that most users can't answer).
[15:45:05] <jayvdb> but maybe I am reading it wrong
[15:45:36] <James_F> jayvdb: http://ee-dashboard.wmflabs.org/graphs/plwiki_ve_hourly_perc_by_ui is the comparative chart, FWIW.
[15:45:48] <MatmaRex> (woot, my wiki hovers at 20%? nice)
[15:45:50] <Maggie_Dennis> Russavia, while I'm only in the office hour to help out with maintenancy stuff, I just wanted to see that I see it more as spiraling upwards. :) Chasing one's tail doesn't lead to progress, and VisualEditor has made a lot of that. I'm not all that comfortable with change myself, but recently have been using it more simply because it's become easier for references. A lot of the improvements...
[15:45:51] <Maggie_Dennis> ...have come from user feedback based on their own uses.
[15:46:00] <Maggie_Dennis> (wanted to say)
[15:46:26] <James_F> mvolz: Yeah, I believe that Analytics have done some experimental analysis using that technique to look at mobile editors. They might be able to apply it to VisualEditor.
[15:46:34] <russavia> Maggie_Dennis: are you familiar with the Antonov An-10?
[15:46:39] <Elitre> MatmaRex: yay Poland! (are you still in the World Cup, though? In that case, forget it...)
[15:46:40] <jayvdb> James_F: maybe you could deploy to anons as a % of registered users with VE enabled. i.e. as the number of registered users enable it, the percentage of anons it is enabled for increases
[15:46:43] <Maggie_Dennis> No, Russavia.
[15:47:11] <Maggie_Dennis> Wikipedia says it's an aircraft, which is a bit out of my area. :)
[15:47:14] <James_F> jayvdb: That… feels odd. Remember that ~95% of registered users never touch a single thing in Special:Preferences.
[15:47:25] <MatmaRex> Elitre: ha, we didn't even make it through the qualifications. the old days of great football in poland are apparently gone :(
[15:47:40] <James_F> jayvdb: Waiting for people to opt-in to VisualEditor on the scale of a wiki is a mug's game; opt-in is for testing and feedback, not deployment.
[15:47:45] <jayvdb> James_F: that chart is much easier to read - thx - but it confirms that pl.wp is avging at most 15%
[15:48:21] <Elitre> (MatmaRex: it's being a tough year for several European teams, apparently.)
[15:48:28] <James_F> jayvdb: Fair. The numbers may have shifted since I last looked (those only show the last month's data).
[15:49:43] <jayvdb> James_F: I meant only that the relationship could be relative. IMO you could switch it on for 5% of anons for every 1% of very active editors who have it enabled
[15:49:47] <russavia> Maggie_Dennis: It was an aircraft designed in the Soviet Union and it had huge problems with crashing, people didn't want to fly on it because it was essentially a deathtrap -- but the government continued to force Aeroflot to use it (probably to keep factories open) -- they tweaked it from time to time, but it kept falling out of the sky -- and it's reputation continued to get worse but still they kept it in operation -- sooner or later co
[15:49:56] <James_F> jayvdb: Oh, yeah, possibly.
[15:50:23] <Elitre> Looks like we only have 10 minutes left :O So grab this last chance to ask James something: if there's not enough time to answer, I'll make sure to send you the answer directly later.
[15:50:25] <Maggie_Dennis> Seems like a pretty big difference between an aircraft that doesn't work and a piece of software that needs user feedback for agile development, Russavia. :)
[15:50:39] <russavia> maggie_dennis: ;)
[15:50:50] <russavia> glad you caught the analogy there though :)
[15:50:53] <jayvdb> if 20% of very active editors have it enabled, it is effectively loved by all and you could force it enabled for the other 80% :P
[15:51:09] <James_F> jayvdb: But switching it on by default for logged-in users (even "just" 5% of them) is a serious step that requires discussion with the community, not just for us to make in IRC in a chat. :-)
[15:51:12] <Maggie_Dennis> Fortunately, if I want to do something in VisualEditor that it can't do, it won't kill me. :/ I say, "Hey! Wouldn't it be cool if..." and they might develop it. :D
[15:51:43] <Maggie_Dennis> (That said, yikes. Note to self: check safety records of airplanes.)
[15:51:46] <jayvdb> James_F: yes, I agree, but I am trying to help you find ways to deploy gradually
[15:51:46] <James_F> jayvdb: I don't think we can ethically say that users with > 100 edits a month get to make all the decisions for others regardless of the impact on others.
[15:52:03] <James_F> jayvdb: It's appreciated.
[15:52:35] <James_F> jayvdb: There's a couple of semi-aborted RfCs on the English Wikipedia about looking at re-enabling it – maybe you should post there?
[15:52:41] <jayvdb> James_F: you made the decision last year that impacted 100% of users :P
[15:53:00] <Elitre> (I don't have stats, but I think I can confirm VE caused no deaths so far. If you know something else, please file it on Bugzilla.)
[15:53:14] <James_F> jayvdb: Major decisions still get made. :-) And yeah, I've said before that it was the wrong call.
[15:53:53] <jayvdb> good to see your ethics have changed since then
[15:55:02] <James_F> jayvdb: The main RfC is https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/2014_RFC and also Pine's https://en.wikipedia.org/wiki/Wikipedia:VisualEditor/VisualEditor_RfC_2014_part_1 alternative (they should probably be merged).
[15:55:17] <Elitre> 5 minutes notice, everyone!
[15:56:21] <jayvdb> thx James. I will take a look at those and get back to you re broader deploy on en.wp
[15:56:30] <James_F> jayvdb: Thanks.
[15:58:49] <Elitre> Thank you all for attending, and thanks James for answering. The log will be posted soon to Meta - for those people who appreciate reading James over and over, like I do. I'm weird, I know.
[15:59:18] <James_F> Thanks everyone. :-)
[16:00:16] <Elitre> We'll see you all in a month here, Saturday July 19th: http://www.timeanddate.com/worldclock/fixedtime.html?hour=21&min=00&sec=0&day=19&month=07&year=2014
[16:00:32] <Elitre> And that's it from us.
[16:00:37] * Elitre waves
[16:01:03] <James_F> Bye everyone.