IRC office hours/Office hours 2011-10-27

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

<poem style="font-family:monospace,Courier;background:#F2F2F2"> [18:59] <Demiurge1000> Just a reminder to everyone that the No Public Flogging rule is not in effect in this channel during meetings. So don't say anything you wouldn't want publicly flogged. [18:59] * Tinz (~Tinz@wikipedia/Tinz) has joined #wikimedia-office [18:59] * ChanServ sets mode: +o StevenW [19:00] <Hurricanefan25> What's "public flogging"? Do you mean no public logging? :S [19:00] <Ironholds> First off, an apology - I got my timezone mixed up. It's technically meant to be in an hour :). I'll be staying on for as long as possible to answer the queries of those who turn up later, but Howie and Fabrice have to head off in about 45 minutes. [19:00] <Demiurge1000> Hurricanefan25: Yes. [19:01] -ChanServ- You have been voiced on #wikimedia-office by StevenW [19:01] * ChanServ sets mode: +v Ironholds [19:01] <+Ironholds> thanks StevenW :) [19:01] <jorm_> Good murder. [19:01] * ChanServ sets mode: +v howief [19:01] * ChanServ sets mode: +v DarTar [19:01] <+DarTar> dankeschön StevenW [19:01] * mindspillage (~kat@wikimedia/KatWalsh/x-0001) has joined #wikimedia-office [19:01] <+Ironholds> So, some introductions. I'm Ironholds, but in the real world people call me Oliver (it's shorter ;p) and I'm the "Community Liason, Product Development". Basically, what this means is that I work to get the community engaging with the foundation, and vice versa, when it comes to developing new bits of mediawiki or changing old bits. [19:02] <@StevenW> no problem :) [19:02] <jorm_> stupid nickserv [19:02] * jorm (~bharris@wikimedia/jorm) Quit (Disconnected by services) [19:02] * jorm_ is now known as jorm [19:03] * tlsv (5ce7e5e3@gateway/web/freenode/ip. has joined #wikimedia-office [19:03] <+Ironholds> we also have fabriceflorin, who is...well, Fabrice Florin :). He's a contractor for the Foundation working on the Article Feedback Tool. His background is in user interface design, and he's designed stuff for Apple, Microsoft and a whole host of other places. He's *incredibly* experienced and great at his job, so please make him feel welcome :) [19:04] <fabriceflorin> Hello everyone, great to meet you, and I am really excited about working with you on this project! [19:04] <WilliamH_UK> ahoy [19:04] <Hurricanefan25> Hey :) [19:05] <jorm> I'm Brandon. I'm a jerk. [19:05] <+Ironholds> there's also howief lurking in the corner over there; that's Howie Fung, who is the WMF's Product Manager. Basically, he oversees the "product stuff" for Wikimedia. I appreciate that might not be helpful :P [19:05] * YE ( has joined #wikimedia-office [19:05] <tlsv> Hi there [19:05] <+Ironholds> hey, tlsv [19:05] * Kingpin13 (~Kingpin13@Wikipedia/Kingpin13) Quit (Ping timeout: 260 seconds) [19:05] <@StevenW> I think it should also be said Fabrice is new to AFT and Wikimedia, so he's not to blame ;) for the first version of the tool. [19:05] <YE> hi [19:05] <Hurricanefan25> Hi YE. [19:06] <sumanah> I'm Sumana Harihareswara, Volunteer Development Coordinator, and will be keeping an eye out for people who want to help out on MediaWiki work via coding, testing, documentation, design, and evangelism. [19:06] <+howief> Hello sumanah [19:06] <fabriceflorin> Yes, and I am new to Wikipedia as well, which makes me a serious newbie -- I look forward to learning from you all ! [19:06] <+howief> any other intros? [19:07] <+Ironholds> I think that's everyone :). So, we're going to talk about the current Article Feedback Tool's succeses and failures, but more importantly what the Foundation is planning to do next. howief, would you like to lead? [19:07] <+howief> Sure [19:07] <+howief> Let me *very quickly* recap what's been done to date [19:07] <+howief> and then we can talk about what we have in mind going fowrard [19:08] * jsalsman (4306b31b@gateway/web/freenode/ip. has joined #wikimedia-office [19:08] <+howief> so the current version of aft went live to the english Wikipedia in July of this year [19:08] * dungodung|away is now known as dungodung [19:08] <+howief> (after a few months of experimentation on the English Wikipedia) [19:08] <+howief> the initial intent of AFT was twofold: [19:09] * Kingpin13 (~Kingpin13@Wikipedia/Kingpin13) has joined #wikimedia-office [19:09] <+howief> 1) to provide an external measurement of article quality that could complement the existing community quality assessment tools [19:09] <jsalsman> is it office hours already? I thought I was early [19:09] <+howief> 2) to test whether feedback on article could be an "on-ramp" for readers into other forms of participation [19:09] <Demiurge1000> But the second of those aims was actually the higher priority one? [19:09] <+howief> well [19:09] <+howief> as it turns out [19:10] <+howief> the first one is very, very complicated [19:10] <+Ironholds> jsalsman: timezone bork. I'll be staying around later as well just to catch people who turn up at the wrong (or right) time. [19:10] <+howief> most assessments of quality are going to conflate two things [19:10] * ChanServ sets mode: +v fabriceflorin [19:10] <+howief> 1) the reader's opinion on the quality of the article [19:10] <+howief> 2) the reader's opinion on the subject of the article [19:10] <jsalsman> Ironholds: thanks; says [19:10] <+Ironholds> jsalsman: indeed. As said, timezone bork ;p [19:11] <+howief> so what we found regarding quality is that with the 4 categories we have in the current version [19:11] <+howief> some correlated well with more objective measure of quality [19:11] <+howief> while others didn't [19:11] * guillom (~guillaume@wikimedia/guillom) has joined #wikimedia-office [19:11] <+howief> participation (the second goal), is a little easier to assess [19:11] <jsalsman> by "more objective" do you mean aggregate editor evaluations? [19:12] <+howief> we found some pretty encouraging results that suggest that AFT could be a good "on-ramp" in converting pure readers into editors [19:12] <+howief> after a user rates an article, we present them with a "call-to-action [19:12] <+howief> " [19:12] <+howief> one of the calls to action was "did you know you can edit this page" [19:13] <+howief> 17% of raters who got that call to action ended up attempting an edit [19:13] <jsalsman> nice [19:13] <+howief> which is a very high number for this type of call to action considering their objective was to rate the article, not to edit [19:13] <Hurricanefan25> howief, the second priority concerns me. The reason I (currently) dislike the AFT is because I think most people rate an article based on their personal opinion of the subject, not the quality of the article. [19:13] <WilliamH_UK> yes, exactly that [19:13] <+howief> there is, of course, the drop-off since not everyone completes the edit [19:13] <+Ironholds> can I ask people to keep general questions about the AFT until the end? We've only got howief and fabriceflorin for another 35 minutes, and we've got a lot to work through :). We've had a couple of general Q&As on the existing tool; this is more meant to be about what we're doing *next*. [19:13] <+Ironholds> If people do have general questions, they're welcome to email me ( and I'll pass them on to the relevant people. [19:13] <+howief> so overall, about 2-3% completed the edit successfully [19:13] <WilliamH_UK> Hurricanefan25 hits the nail on the head [19:14] <+howief> Hurricanefan25: yes, it's very difficult to disambiguate the two [19:14] <+howief> which is why we're heading in another direction [19:14] <jsalsman> Ironholds: is there a wiki page where questions can be posted? [19:14] <+DarTar> Hurricanefan25: WilliamH_UK - that's certainly true for trending/very popular articles [19:14] <YE> I think it should be removed for non-vital articles, as they often have 0 page ratings [19:14] <+Ironholds> (one moment) [19:14] <+DarTar> (which attract polarized opinions) [19:14] <+howief> based on the analysis we've done and discussions with the community during the first phase [19:15] <+howief> we'll be focused more on finding ways to have readers contribute productively to the building of the encyclopedia [19:15] <+Ironholds> jsalsman: sure; the talkpage works [19:15] <jsalsman> Hurricanefan25: the separate rating questions should be able to minimize the effects of opinions of the subject versus article quality, and we could check that if the data were public [19:15] <+howief> we'll still be looking at quality, but we'll be testing various methods of assessing quality implicitly vs. explicitly [19:15] <+howief> in an effort to avoid the "i dislike justin beiber so i'm going to rate this article poorly" problem [19:15] <+howief> so that's the basic recap [19:16] <+DarTar> jsalsman: all data is public [19:16] <jsalsman> great! [19:16] <+howief> i'll turn it to fabrice to talk about what we have in mind for the next version [19:16] <+fabriceflorin> In the next few weeks, we plan to create and test a couple different versions of a simple feedback form for AFT V5 [19:16] <+Ironholds> (for context - that's Version 5. We're on Version 4 at the mo) [19:16] <WilliamH_UK> It's also true for controversial articles or fringe subjects with a particular following. My content edits concern controversial stuff like Holocaust denial, and the idea that Holocaust deniers reading the article can rate how trustworthy or objective it is, when their opinion on the issue is editorially disappointing [19:17] <+fabriceflorin> Yes, good point, Ironholds. You can see preliminary designs for these feedback forms on our AFT V5 page: [19:17] <+fabriceflorin> [19:17] <+Ironholds> WilliamH_UK: sure. Again, can you keep objections until we've set out what the new ideas are? They may render the objections moot. [19:17] <sumanah> fabriceflorin: quick question for context -- remind us about how many developers/designers are working on this? [19:17] <WilliamH_UK> sure [19:17] <+fabriceflorin> This first batch of feedback forms will all have a comments feature, which many people told us was important to them [19:17] <YE> There fine for high-viewed articles [19:18] <+howief> sumanah: we're working with an outside development firm to do this, but let's keep implementation questions to the end. I'd like for us to focus on the feature ideas and direction first [19:18] <+fabriceflorin> We hope they will encourage readers to provide constructive feedback, rather than judgmental ratings [19:18] <sumanah> ok howief, thanks [19:18] * Tiptoety (~Tiptoety@wikipedia/Tiptoety) has joined #wikimedia-office [19:18] <YE> but for pages that get less than 100 views a day, they are useless imo [19:18] <+fabriceflorin> For example, here is a wireframe for one of the feedback forms we are now working on - Option 1: [19:18] <+fabriceflorin> [19:18] <+fabriceflorin> This first option will ask if you "found what you were looking for", with 'yes' or 'no' buttons and a comment field. [19:19] <Hurricanefan25> I dislike that one - though it provides more essential feedback [19:19] <+fabriceflorin> This approach was recommended by community members and invites suggestions for improvement rather than ratings [19:19] <YE> they often get 0 page ratings [19:19] <+Ironholds> Hurricanefan25: can you keep feedback until we've set out all the options? [19:19] <+Ironholds> YE: this is true. However, again, please keep comments... :) [19:19] <+fabriceflorin> It can also provide implicit measurements of quality that could be as effective as explicit ratings in the current form. [19:20] <+fabriceflorin> What do you think of this first feedback form? what do you like most? least? how could we improve it? [19:20] * jsalsman thinks it is a little odd to shush questions when the announced meeting time was incorrect [19:20] <+howief> well let's just chat about this feedback form [19:20] <Tinz> will the reader comments be stored at a separate wikipage or at the talkpage? will admins be able to edit/delete inappropiate comments [19:20] <+howief> Hurricanefan25: what do you dislike about it? [19:21] <+fabriceflorin> Glad you asked Tinz, we have a feedback page that will let you do that. [19:21] <+fabriceflorin> We will show you a preliminary design for that feedback page in a moment. For now, what do you guys think of this first feedback form? [19:22] <Hurricanefan25> howief: To be specific, it doesn't have the same requirement for providing feedback based on type. Someone could type "lololol he sux loldong" into the field and that would be utterly unhelpful; though with the stars, the same thing happens in a way, by rating everything one star [19:22] <+fabriceflorin> [19:23] <+Ironholds> Hurricanefan25: if you look at fabriceflorin's link (just above this comment) you'll see unhelpful comments can be removed or hidden. I think they also plan on running them through the abuse filter [19:23] * @StevenW (~stevenwal@wikimedia/steven-walling) Quit (Read error: Connection reset by peer) [19:23] * StevenW (~stevenwal@wikimedia/steven-walling) has joined #wikimedia-office [19:23] <jorm> local wifi just dropped. [19:23] <+Ironholds> oh lord :(. Okay, they'll be back in a sec, hopefully :p [19:24] <Hurricanefan25> howief: However, this problem could possibly be solved in a two-step solution: 1) Require a rating for the stars and 2) require text feedback based on each rating, like a little bit of text that says "provide specific feedback" - and if no specific feedback is given, use a little JavaScript or something to say "You haven't provided any specific text feedback. Would you like to continue?" or... [19:24] <Hurricanefan25> ...something like that. [19:24] <+Ironholds> Hurricanefan25: I think fabriceflorin is about to bring up the second feedback tool, which is just text feedback. That might solve the issue :) [19:24] * howief_ (~howiefung@ has joined #wikimedia-office [19:24] <+DarTar> Hurricanefan25: we're going to a/b test different options and measure the noise we get for each of them [19:24] * tlsv (5ce7e5e3@gateway/web/freenode/ip. Quit (Ping timeout: 265 seconds) [19:24] <+Ironholds> welcome back, howief! [19:24] <howief_> that sucked [19:24] <howief_> so yeah [19:24] <jorm> Any comment field will have a signal to noise ratio. [19:24] <jorm> So we need to test different types. [19:24] <Tinz> so technically, the comments wont be stored in a usual mediawiki page such as the discussion page? [19:25] <+fabriceflorin> Yes, Tinz, they will be visible on another page. More on that later. [19:25] <+fabriceflorin> Here is another feedback form we would like to test in coming weeks - Option 2: [19:25] <+fabriceflorin> [19:25] <jsalsman> I think the new design is far superior. I absolutely love it. However, I wish you would please combine Option 1 and Option 2 [19:25] <+fabriceflorin> This second option simply asks you to add a suggestion, ask a question, report a problem or give praise to the authors. [19:25] <howief_> how would you think of combining the two? [19:25] <Hurricanefan25> Ditto jsalsman [19:26] <jsalsman> I think the Yes/No question "Did you find what you were looking for?" is really useful for aggregation [19:26] <howief_> and how would that help the editing community? [19:26] <howief_> i'm not disagreeing at all [19:26] <+fabriceflorin> We had an earlier design that combined the two, but are aiming for maximum simplicity. Less is more! [19:26] <Hurricanefan25> Ditto, though I only think there should be "suggestion" or "criticism" on option two [19:26] <howief_> in fact [19:26] <howief_> i think it's a good idea [19:26] <howief_> i'm just trying to get a sense for how the editing community would use this feedback [19:26] <+DarTar> and less is more easily measurable :p [19:26] <jsalsman> and I would love to see the textarea be prefilled with disappearing "How can we improve...?" instead of "Describe your suggestion" [19:27] <howief_> the more we know about how this type of feedback is used, the better informed our development team can be [19:27] <Bensin> What's the difference between "Suggestions" and "Problem"? [19:27] * +howief (~howiefung@ Quit (Ping timeout: 260 seconds) [19:27] * howief_ is now known as howief [19:27] <+fabriceflorin> Hurricane, note that the second form starts with 'suggestion' by default. Note that it doesn't ask you if you found what you were looking for, and focuses primarily on your comment. [19:27] <jsalsman> but the suggestion/question/problem/praise choices should be checkboxes, not radio buttons [19:27] <+Ironholds> Bensin: I imagine suggestion is "you might want to add in X" and "problem" is "there's vandalism/you got this bit horribly horribly wrong" [19:27] * ChanServ sets mode: +v howief [19:28] <jsalsman> it is possible to have a suggestion, question, problem, and praise all at once [19:28] <+howief> jsalsman: i'm not sure [19:28] <+howief> that may be something we test [19:28] <+fabriceflorin> jsalman, we like to think of these options more like radio butons than checkboxes, but will take your suggestion under consideration [19:28] <+howief> the simplest interaction from a reader standpoint might be to just enable the reader to select one [19:28] <jsalsman> instead of spending money making it look all Web 2.5, why don't you use regular HTML v3 forms which will render on people's phones? [19:29] <+howief> but this is something we can play around with [19:29] * Hurricanefan25 (~chatzilla@wikipedia/Hurricanefan25) Quit (Quit: ChatZilla 0.9.87 [Firefox 4.0.1/20110413222027]) [19:30] <+fabriceflorin> If you select one of the 'tabs' in option 2 (suggestion, question, problem or praise), you can get a separate prompt and possibly different calls to action after you comment. [19:30] * Philippe (~Philippe@ has joined #wikimedia-office [19:30] * eiaway (~chatzilla@wikimedia/effeietsanders) has left #wikimedia-office [19:30] <jsalsman> because mobile devices are soon going to outnumber client page loads from traditional desktops. You don't want to get caught spending tons of time on javascript that won't run or run well on mobiles instead of basic HTML forms [19:30] <+howief> jsalsman: yup [19:30] <jorm> So, how this will interect with mobile is something I'm directly involved with and super-focused on. [19:30] * RoanKattouw is now known as RoanKattouw_away [19:30] <jsalsman> how many mobile clients do you test on? [19:31] <jorm> A couple hundred, I believe. [19:31] <jsalsman> wow! [19:31] <jorm> You'd have to ask Tomasz for the exact numbers. [19:31] <+howief> we should have another one of these for mobile [19:31] <jorm> however, I think mobile is outside of the scope of this conversation right now. [19:31] <jsalsman> that's way up from last year [19:31] <+howief> as jsalsman points out, mobile is going to be more and more important [19:31] * YE ( has left #wikimedia-office [19:32] * tlsv (5ce7e5e3@gateway/web/freenode/ip. has joined #wikimedia-office [19:32] <+fabriceflorin> Which version of the feedback form do you prefer? option 1? (did you find what you were looking for?) or option 2? (suggestion, question, problem, or praise). [19:32] <sumanah> people who want to talk about mobile stuff, we welcome you in #wikimedia-mobile [19:32] * Philippe (~Philippe@ Quit (Client Quit) [19:32] <Bensin> What's the purpose of the question "Did you find what you are looking for"? [19:32] * Zidanie5 (~Zidanie5@wikipedia/zidanie5) has joined #wikimedia-office [19:33] <+howief> Bensin: during a previous office hours on this subject [19:33] <+DarTar> Bensin: the point is to make the feedback actionable [19:33] <Tinz> i prefer option 1, less complicated to use [19:33] <+howief> i asked the group what they'd like to know from readers [19:33] <+fabriceflorin> Please tell us why you prefer one option over the other? [19:33] <+howief> what readers came to wikipedia looking for came up as something that might be useful to help in the development of articles [19:33] <+howief> e.g., what's missing from the article [19:34] <jsalsman> I would love to see plain vanilla traditional radio buttons for "Did you find what you were looking for? Yes ( ) No ( )" followed by four checkboxes "Suggestion [ ], Question [ ], Praise [ ], Problem [ ]" and then a text area that says "How can we improve this article?" either inside it or above it. And please test it with Javascript and image loading turned all the way off [19:34] <jorm> This feature is more than likely going to be a javascript only feature. [19:34] <jsalsman> because I can't even get the reference desks to load most of the time without both images and javascript off on my 2.5 year old phonw [19:34] <jsalsman> phone* [19:35] <jorm> We aren't committing to that, or any other direction. [19:35] <jsalsman> why javascript only? [19:35] <Bensin> DarTar: Actionable? Action from whom? To do what? [19:35] <+howief> jsalsman: we haven't figure out the details to that yet. one of the things we will try to do is to have the feature "degrade gracefully" for non-js [19:36] <+fabriceflorin> Thanks for your thoughtful suggestions about the feedback forms. After this chat, please add your comments on this talk page on Wikipedia-EN: [19:36] <+fabriceflorin> [19:36] <+Ironholds> Bensin: actionable = so the feedback can be used by editors to make improvements [19:36] <+howief> meaning raters can still use the feature, but it may not behave in exactly the same way [19:36] <+howief> but we'll need to figure out when we can roll in non-js support [19:36] <+Ironholds> so, if the readers say "this article is a 3 for completeness", that's not necessarily actionable. Editors might not know what is missing. [19:36] <+fabriceflorin> During the month of November, we will build a couple versions of this feedback form, then plan to test them in December [19:36] <+howief> as you can imagine, we're developing quite a long laundry list, and things need to be prioritized [19:36] <+fabriceflorin> In coming weeks, we will also develop a new feedback page, where all these comments could be listed for each article. [19:36] <+Ironholds> if they say "you should add the date of birth", that's actionable - an editor can go "huh, that's a good idea" and react. [19:37] <+fabriceflorin> Here is a very preliminary example of what this feedback page could look like: [19:37] <+fabriceflorin> [19:37] <jsalsman> if you build it without javascript dependencies to begin with, you save money and time and don't have to worry about graceful degradation [19:37] <+DarTar> actionable for editors, as a first specification for the feedback collected via the tool as opposed to feedback-less ratings as is currently the case [19:37] <+fabriceflorin> All the feedback posts for a given article would be listed on this page, where they can be moderated by editors and administrators. [19:37] <+fabriceflorin> Editors would have the option to feature the best feedback at the top of this page -- or hide offensive comments [19:38] <+fabriceflorin> We're also considering simple tools to let community report abuse, as well as vote feedback posts up or down. [19:38] <jorm> Let's move on from questions about implementation. [19:38] <Bensin> Ironholds: Ahhh. I think I understand... To achieve that it should be phrased much clearer. [19:38] <jsalsman> a mobile phone today with low signal strength is like a 14.4kbps modem in 1994, with the same rationale for turning off images and javascript [19:38] <+Ironholds> Bensin: sure :) [19:38] <jorm> jsalsman: please let's move on from this. [19:39] <+fabriceflorin> What do you think of this preliminary design for the feedback page? what do you like most? least? how could we improve it? [19:39] <Bensin> I don't want to commit to any of these suggestions without thinking about it first. But I like the move away from ratings. [19:39] <+Ironholds> Bensin: that's cool :). So, we're going to be discussing it at the talkpage of - when you've had a think, please comment there! Your feedback is appreciated :) [19:39] <jsalsman> jorm: do you think I've been inaccurate? [19:40] <+fabriceflorin> Thanks, Bensin, I am glad that you like the shift away from ratings, which sounds right for me as well. [19:40] <+howief> jsalsman: we'd like to cover the feedback page [19:40] <+howief> it's an important page for the community since this is where the comments go [19:40] <+fabriceflorin> And I appreciate that you will be thinking more about this. Your feedback means a lot to us. [19:40] <+howief> maybe we can talk about js support later? [19:40] <jorm> I think that we're too early in the process to have a discussion about technical implementation, and it is only going to muddy the issue. So let's move on. [19:41] <+fabriceflorin> What do you think about the feedback page? How could we make this preliminary design better? [19:41] <+fabriceflorin> [19:41] <jsalsman> are you proposing to put a scrollbar text area inside a page? [19:42] <jsalsman> and if so, are there any reason that the text can't be placed on the page without a scrolling subwindow? [19:42] <jorm> This is only a wireframe prototype. We're not in the design phase. [19:42] <jsalsman> when PubMed did that to their limit checkboxes, I couldn't get to those on my phone either [19:43] <jorm> Please don't get hung up on details in these screens; just focus on the intents. [19:43] <Bensin> fabriceflorin: I like the feedback page! I'd like an option where I as an editor can mark a reader comment as "taken care of". [19:43] <jsalsman> well do you want feedback or not? [19:43] <+Ironholds> jsalsman: we do; at the moment we just want to know what you think of the ideas being bandied about; the specifics come later. [19:44] <jsalsman> someone has to stick up for mobile users at the front end of the process or it will be a mess trying to retrofit them [19:44] <+howief> jsalsman: agreed [19:44] <+fabriceflorin> Bensin, that is a GREAT IDEA! We'll think of a way to identify which reader suggestions have been addressed. [19:44] <+howief> i think we should do a mobile-specific office hours with the mobile team [19:44] <Tinz> cant say much about design questions, but I am a bit sceptical that this might turn into a parallel discussion page, just not in mediawiki format, which could lead to a lot of confusion [19:44] <+howief> get tomasz, phil, and patrick on the discussion (dev manager, product manager, and engineer for mobile) [19:44] * kellerkind71 ( has joined #wikimedia-office [19:45] <+howief> Tinz: that's a very important consideration [19:45] * pchang (~pchang@ has joined #wikimedia-office [19:45] <+howief> we can't have two parallel discussion pages [19:45] * pchang (~pchang@ Quit (Remote host closed the connection) [19:45] <+howief> so when we get into the detailed design and user interaction [19:45] * pchang (~pchang@ has joined #wikimedia-office [19:45] <+howief> we need to make sure the concepts are cleanly separated [19:45] <+howief> if we can't cleanly separate the concepts, we'll have to figure out another way of manning the comments [19:45] * Weresplchqrs (522bc8d8@gateway/web/freenode/ip. has joined #wikimedia-office [19:45] * jsalsman_ (4306b31b@gateway/web/freenode/ip. has joined #wikimedia-office [19:46] <+fabriceflorin> Good point, Tinz. We would like the moderation tools to eventually let you promote the most useful feedback to the talk page, where editors can act on them. This could become a "work-list" for editors working on an article, as Jorm has wisely suggested. [19:46] <+howief> but i agree that this should not turn into a parallel discussion page [19:46] <jsalsman_> sorry got disconnected [19:46] <+fabriceflorin> Thanks for these great ideas! We'll take them in consideration as we refine our designs for this new page. [19:46] <Bensin> fabriceflorin: Also, how about making the feedback page visible on the article talk page. The same way an articles "To-do"-list is visible on the talk page. [19:46] <+fabriceflorin> These forms and other variations will be released and tested on a subset of English Wikipedia pages starting in December 2011. [19:46] <+fabriceflorin> We will collect detailed metrics for each of these feedback forms, to measure their effectiveness in engaging readers and assessing article quality [19:46] <jsalsman_> are there any objections to doing the first pass of usability testing on phones with javascript and image loading off? [19:46] <+fabriceflorin> We will also track overall satisfaction by readers and editors, through meetings like these, as well as surveys and other channels. [19:47] <+fabriceflorin> Our development process will be based on short milestones, open to the public, and we will involve you at each step of the way [19:47] <+fabriceflorin> We have also engaged an excellent development studio to help build some of these features: OmniTI in Maryland [19:47] <+fabriceflorin> They are experts in this kind of work, their software is used by hundreds of millions of users, and they have won many awards. Check their site: [19:47] <+fabriceflorin> [19:47] <+howief> jsalsman: i can't promise that, but we'll certainly try. but let's do this in concert with the mobile folks [19:48] * Zidanie5 (~Zidanie5@wikipedia/zidanie5) has left #wikimedia-office [19:48] <+fabriceflorin> Thanks for all your great suggestions! They will really help guide our next steps. [19:48] * tlsv (5ce7e5e3@gateway/web/freenode/ip. Quit (Ping timeout: 265 seconds) [19:48] * jsalsman (4306b31b@gateway/web/freenode/ip. Quit (Ping timeout: 265 seconds) [19:49] <+fabriceflorin> Bensin, good point about making a link to the feedback page visible on the article page. We are considering ways to do this, besides the link that is already in the feedback form. [19:50] <Weresplchqrs> How do we avoid this becoming a "a "work-list" for editors working on an article" and steer the commenters into actually improvng the article instead of them telling existing editors what to do? [19:50] * kellerkind71 ( Quit (Remote host closed the connection) [19:50] <Bensin> fabriceflorin: No, I was suggesting that the _content itself_ on the feedback page is visible on the article talk page. May be good. May not be. Just consider it. [19:50] * jorm (~bharris@wikimedia/jorm) Quit (Read error: Connection reset by peer) [19:51] <+howief> hey folks i gotta run [19:51] * jorm (~bharris@wikimedia/jorm) has joined #wikimedia-office [19:51] <+howief> i think Ironholds is going to set up another one of these soon so see you all then [19:51] <+DarTar> same here, happy to continue the conversation on the wiki [19:51] <+fabriceflorin> Good point, Weresplchqrs. One of the options we are testing is a simple call to action to 'Edit this page'. [19:51] * +howief (~howiefung@ Quit (Quit: howief) [19:51] * WilliamH_UK waves to Weresplchqrs [19:51] * StevenW (~stevenwal@wikimedia/steven-walling) Quit (Quit: StevenW) [19:51] * +DarTar (~DarTar@wikimedia/DarTar) Quit (Quit: DarTar) [19:51] * jsalsman_ is now known as jsalsman [19:52] <+fabriceflorin> Ah, Bensin, thanks for the recommendation to feature some of the feedback page content on the article page. We may be able to feature the best comments to the side of the feedback page. [19:52] <+fabriceflorin> You guys are awesome! Some really good ideas came out of this conversation. I look forward to reconnecting in a week. [19:53] <+fabriceflorin> For now, you can read more about our detailed plans for AFT V5 here: [19:53] <+fabriceflorin> [19:53] <+fabriceflorin> With your help, we can give a voice to Wikipedia readers, get them to participate more actively, and engage them to become editors over time. Onward! [19:54] <Bensin> fabriceflorin: No, still not what I mean :-). I mean that the _entire feedback page_ is visible on the talk page. Same as the Template:To do [19:54] <Weresplchqrs> Waves to Will, Thanks Fabrice, good to hear you are looking at that. It is important when we measure the results to not at the number of edits but whether these new editors actually do start fixing things or they just continue to critiqu. [19:54] <jsalsman> fabriceflorin: is on your watchlist? I'd prefer to comment there [19:55] <+Ironholds> jsalsman: it will be; we're (at least initially) going to try and keep the discussion on enwiki, I think, just because it's easier for the vast majority of editors not to be flipping between projects. [19:55] <+Ironholds> I understand this might put you in particular in an awkward situation, so feel free to comment on the mediawiki talkpage :) [19:55] <+Ironholds> if everyone else could stick to enwiki, at least initially, I'd be very grateful. [19:56] <jsalsman> is the AFT currently only in use on enwiki? [19:56] <+Ironholds> I *believe* so - fabriceflorin? [19:56] <+fabriceflorin> Yes, the best place to add your comments is on this talk page on Wikipedia-EN: [19:56] <+fabriceflorin> [19:57] <jsalsman> Perhaps it would be best if I corresponded with Theo and Cipirian [19:57] <+fabriceflorin> Yes, the current AFT is only on the English encyclopedia during this experimental phase. [19:57] <jsalsman> Ciprian* [19:58] * WilliamH_UK (WilliamH_U@Wikipedia/WilliamH) has left #wikimedia-office [19:58] <+fabriceflorin> Thanks, everyone. It's been great! Let's talk again in a week. [19:59] <+fabriceflorin> And thanks to Ironholds for organizing this event, much appreciated! Bye for now. [19:59] <Demiurge1000> thanks all [19:59] <+Ironholds> Okay, so, the devs are going to be looking for feedback at a lot of stages in the process: first off, throughout the whole thing, we've got (as Fabrice says) which briefly explains all the bits and bobs that are being worked on. [19:59] * +fabriceflorin (~fabricefl@ Quit (Quit: fabriceflorin) [19:59] <+Ironholds> The talkpage is going to be the main venue, at least initially, so please post any ideas, suggestions or comments on the work there :). If (as jsalsman is) you don't want to comment there, you can do so at the mediawiki page [19:59] <jsalsman> thanks [19:59] <+Ironholds> but I'd ask people to stay on enwiki at least initially [20:00] <+Ironholds> (unless they have a good reason) [20:00] <+Ironholds> If you look at the bottom of that page, there's a schedule; the bars marked in green are ones where we're really going to be engaging editors. [20:00] <Bensin> Is it bugging devs when editors come to their turf? [20:00] <jsalsman> so, how about the automated reading level assessment you've been working with, Ironholds, should that be part of the article feedback statistics? [20:00] <+Ironholds> so the first one is 3 November, when the first mockups come out :) [20:01] * memethuzla (574a4934@gateway/web/freenode/ip. has joined #wikimedia-office [20:01] <Demiurge1000> Bensin: No, but vice versa can sometimes be the case :) [20:01] * Jan_eissfeldt ( has joined #wikimedia-office [20:01] <+Ironholds> jsalsman: I'd say "no"; both the SMOG Index and Gunning-Fog systems have idiosyncracies which make them unhelpful for wikipedia. We actually had a tool that allowed for automated assessment - it was shut down because the ratings just weren't very useful [20:02] <jsalsman> has anyone tried them without out of vocabulary scores? That triggers off spelling problems, too [20:02] <Bensin> I was a little bugged when devs rarely answered questions about AFT on the Village pump and when editors had to go to the MediaWiki page to get answers. Not good. [20:02] <+Ironholds> anyway. So, 3 November, the first mockups are released - we'd like your perspectives then particularly. You'll also see the first feedback form designs come out on 10 November, and the first feedback *page* on the 23rd. Those are the most immediate dates. [20:02] <+Ironholds> Bensin: understandable; the devs will be engaging a lot more on the enwiki page in this design process, even if I have to poke them with a stick to do so :P [20:02] <jsalsman> i.e., all the automated scoring systems I've seen score obscure words worse than you probably would want for an encyclopedia [20:02] <+Ironholds> jsalsman: well, no [20:03] <jsalsman> ...same as misspelled words [20:03] <+Ironholds> so, SMOG Index and Gunning-Fog rate based on things like number of polysyllabic words, number of words in a sentence, number of punctuation elements in a sentence. [20:03] <+Ironholds> those are the most normal readability algorithms. The problem is they don't produce very accurate results on Wikipedia, because "long word" is not the same as "complicated word". One of the words SMOG jumps at, because it's Wikipedia ;p [20:04] * YairRand (63ee0d28@gateway/web/freenode/ip. has joined #wikimedia-office [20:04] <Demiurge1000> And pollysyllabic. [20:04] <Demiurge1000> And readability, and... [20:04] <jsalsman> my point, we ought to be able to keep sentence length and similar measures and ignore out-of-vocabulary (and misspelled) words [20:04] <memethuzla> Wikipaedia? [20:04] <+Ironholds> but this is getting a bit off base. If anyone has feedback or comments you want to jot down or think about, is going to be the main venue. Post them there :) [20:05] <+Ironholds> I'll be sticking around in this channel for another hour or so due to the problems we had with timing - I don't want to leave late arrivals to miss everything - so if you have questions about community engagement (which is my area), ask away. [20:05] <jsalsman> where will the transcript of the previous hour be posted? [20:06] <+Ironholds> the usual location; I'm just copying them into a text document and culling them for irrelevant bits (joins, parts, that sort of thing) as we speak [20:07] <Bensin> Ironholds: To me the problem with AFT has been twofold. 1) The tool itself (which is now being addressed and seems to improve for the better) 2) The process of how it has been implemented. Can we talk about this process? [20:07] * fabriceflorin (~fabricefl@ has joined #wikimedia-office [20:07] <+Ironholds> so that's going to be - it'll have an entry under "office hour logs" [20:07] <+Ironholds> Bensin: sure, go ahead :) [20:08] <+Ironholds> Bensin: sorry - wires crossed - do you want me to talk about the new process, or do you want to give opinions on the old implementation? [20:09] <Bensin> The AFT has a HUGE impact on the entire community and readers of a wiki. And I the way the foundation just switched it on without a clear community consensus... That's what I want to talk about. [20:09] * Maryana (~Maryana@ Quit (Ping timeout: 276 seconds) [20:09] <+Ironholds> okay, that's fair enough. Can you explain what the impact on editors was? Just so I know we're on the same page [20:11] <Bensin> The tool is visible on every single page on the wiki. For the readers: They see it, are encouraged to rate. For the editors: The are (in a way) expected to take some sort of responsibility for the articles and improve them. [20:11] * +Ironholds nods [20:13] <Bensin> How the devs at the foundations just decided to switch it on indicates a shift on how the foundation sees itself. What role it has towards the community. [20:13] <+Ironholds> It wasn't implemented with clear community consensus, no - generally tech changes aren't. It's just not practical most of the time. There were some attempts to get the community involved, some productive, some not, and I don't want to revisit the wrongs or rights of the Foundation's approach in relation to community comments just because well...I wasn't there :). [20:13] <+Ironholds> and I don't want to be putting words in staffers' mouths. [20:14] <Bensin> I understand. [20:14] <Weresplchqrs> As well as furthering the trend to critiquing rather than improving articles, the AFT took development and testing resource that could have done useful work. [20:14] <+Ironholds> it wasn't just an arbitrary switch on, though; there were attempts to get feedback and notify people before it was turned on. Whether this was done well or badly is another matter, though, which like I say, I don't feel comfortable revisiting. [20:14] * guillom (~guillaume@wikimedia/guillom) has left #wikimedia-office ("Quitte") [20:14] * Tiptoety is now known as Tiptoety|busy [20:15] <YairRand> I find it odd that it was enabled specifically on enwikipedia. not sure why that was done. [20:15] <+Ironholds> I can't promise that the developers are going to wait for a 300-person majority or whatever this time, either. What I *can* say is that I'm going to be working as hard as possible to make them take legitimate concerns into account, and as Fabrice's reactions to your comments has shown, this isn't likely to be hard. They are trying to improve. [20:15] <+Ironholds> And, practically speaking, if big chunks of the community hate it, the devs would be mad not to give it some attention [20:15] <Bensin> Ironholds: Well I can say I voiced my opinion and was basically told "this is going to happen". [20:16] <+Ironholds> Bensin: sorry to hear that :(. Fabrice is taking a lot of your comments on board - I heard him over skype reacting to one of your suggestions with this booming American "GREAT IDEA!" :P [20:16] <+Ironholds> so hopefully the process will be done a lot better this time. I'm going to try and make sure it is. [20:16] <memethuzla> Even now there isn't a "rate this rating device" box. No pointer to dev page. [20:17] <+Ironholds> in relation to general community reactions, actually, I would like to say one thing (but I'll reply to Weresplchqrs first) [20:17] <Bensin> I have little doubt the AFT will improve. Everything points to that. And that is good. [20:17] <+Ironholds> Weresplchqrs: that's been taken into account. The "call to action" messages (inviting readers to get involved and edit) are going to be a key part of the new system. I think the uptake rate was around 17 percent last time (17 percent of people who read the message inviting them to edit clicked through) [20:17] <+Ironholds> the challenge is increasing that, and increasing the number of people who *succeed* in making an edit. [20:18] <+Ironholds> (okay, I'll be back in two minutes, then I'll respond re community reactions) [20:20] <+Ironholds> righto, back [20:20] <Weresplchqrs> But the process we should be thinking of is not how we implement the Articcle Feedback Tool. The process we should be looking at is how do we decide which of the many possible changes to the wiki do we prioritise. Currently this seems to be an internal process at the Foundation, with things that might interest readers given a priority and things that would make things better for editors given minimal priority. [20:20] <+Ironholds> so, actually the reaction from both readers and editors to the old version has been quite good - if you look at you'll see that a vast majority think the tool is good. Obviously this has self-selection bias (it was a survey, after all) but I'd still say, generally, that it's seen as A Good Thing. [20:21] <+Ironholds> but (and it's a big but, the sort you could write a 90s hip-hop song about) [20:21] <+Ironholds> Just because the AFT was popular with most people doesn't mean the people who didn't like it don't have good points, and it doesn't mean we can't improve it. That's why we're revisiting the design. I can't promise we'll never implement except with consensus, because tech doesn't work like that, but I can promise we'll take concerns into account to try and build the best AFT possible. [20:22] <Weresplchqrs> How much IT resource is going into fixing the problem with Catalot and cats in templates? [20:22] <+Ironholds> Weresplchqrs: in regards to, sorry..? [20:23] <jsalsman> supposedly there is a Foundation process to set development priorities, and it seems okay, but very flawed too. Last year offline editing made the top 20 strategic goals, but it was soon withdrawn as "clunky." sumanah, do you know more about how that happened? [20:23] <+Ironholds> I know Brion is (or was?) working on fixing template syntax, Weresplchqrs. I can't really comment on general tech priorities, though, it's not my area. In regards to your low priority/high priority point, the Foundation is doing editor-centred stuff just as prominently. [20:24] <+Ironholds> as an example - at the moment, I'm working on the new AFT thing half the time. The other half, I'm spending getting feedback on a new Special:NewPages interface: something entirely editor-centric that should help (in some small way) make the editing experience easier. [20:24] * sumanah is in another meeting, will respond in like 30 min [20:25] <+Ironholds> to move slightly back on track - does anyone have comments about the AFT process specifically? Or are people interested in just asking general queries? Warning: I probably won't be able to answer large chunks of those :) [20:25] * Odisha1 (Odisha1@ has joined #wikimedia-office [20:25] * Odisha1 (Odisha1@ Quit (Changing host) [20:25] * Odisha1 (Odisha1@wikimedia/odisha1) has joined #wikimedia-office [20:27] <Bensin> Ironholds: I hear you. But those numbers were pulled after the tool was implemented. I am, at this point, not saying the tool should be removed (or stay). I am saying that I want the foundation's main purpose to be to support the community. Mostly it does, but with the AFT something went really wrong and the attitude just broke my heart a bit. [20:27] * Weresplchqrs (522bc8d8@gateway/web/freenode/ip. Quit (Ping timeout: 265 seconds) [20:27] <+Ironholds> Bensin: that's fair enough. I think it's a matter of different perspectives [20:28] <Bensin> I think it's a matter of the foundation not having a perspective at all :-) [20:28] <+Ironholds> so, editors (by and large) see tech's job as supporting and developing the tools they need. That's fair enough - it is and should be a major priority for devs to make the editing experience better. [20:28] <jsalsman> I think the new designs are really good [20:28] <+Ironholds> jsalsman: excellent! :) [20:28] * Rock_drum (~Rock_drum@wikimedia/Rock-drum) has left #wikimedia-office [20:28] <jsalsman> but I hope they use ordinary forms, and combine the two options [20:29] <+Ironholds> but devs aren't just responsible for editors, they're responsible for readers as well. So the "community", from the Foundation's purpose, is editors, readers, researchers, and everyone else. [20:29] <jsalsman> there is no reason that someone should be forced to enter two comments if they have both praise and a question, for example [20:29] <+Ironholds> There's this foundation word "stakeholders" (which, to be honest, makes me want to drive a nail through my forehead) that they use to describe who they're responsible to. And it's a good word, even though I hate it - they're responsible to everyone who has a stake in the projects. And readers are one of those groups. [20:29] <jsalsman> the "Did you find what you were looking for?" is a great yes/no question, and the primary measure of the utility of reference works [20:30] <+Ironholds> jsalsman: I'm scribbling your comments down now, and will throw them over to the devs. I've got your email address written down somewhere - is that the easiest way to contact you with their responses? [20:30] <jsalsman> [20:30] <+Ironholds> jsalsman: cool beans. Thanks :). [20:31] <Bensin> Ironholds: (as for the question I'd prefer "Did this article contain the information you were looking for?") [20:31] <+Ironholds> Bensin: also written down :). As said, the talkpage is probably the best way to communicate; people will be watching it, and it lets all the interested parties see what everyone else is saying too. [20:31] <Bensin> Gotcha [20:32] <jsalsman> if we need to pick between the "Did you find what you were looking for? Yes/No" and the "Suggestion/Question/Problem/Praise" selection, whether it's a radio button or multiple response selection, I'd go with the former, but I really want both with the checkboxes [20:32] <+Ironholds> jsalsman: noted :). I'll scribble it down. Are you planning on engaging on mediawiki too? [20:32] <memethuzla> I had suggested "Are important relevant facts missing? (Please sign in, & share your knowledge.)" [20:33] <jsalsman> there is even a perfectly legitimate logical and user experience argument for allowing both "Yes" and "No" to be selected. As much as electrical engineers don't like to think about it, "yes and no" is a perfectly legitimate answer to the question [20:33] <jsalsman> yes [20:33] <jsalsman> I [20:33] <jsalsman> I'll comment on the MediaWiki page [20:33] <+Ironholds> jsalsman: yeah, OR functions are for circuit design, not for UX work [20:33] <+Ironholds> (I can't believe I just made an electrical engineering comment. Finally, that GCSE in Electronics makes itself useful) [20:35] <+Ironholds> memethuzla: noted. Would you also like to post your suggestion on the talkpage? Like I said, it's going to be the main venue - we're trying to get people to go there as a first port of call to keep the discussions transparent. [20:35] <+Ironholds> (which is also one of the reasons I'm trying to keep stuff on enwiki, at least initially, unless there's a good reason not to) [20:35] <memethuzla> yes, it's already there [20:35] <+Ironholds> memethuzla: ? [20:37] <+Ironholds> or did you mean the general AFT talkpage? [20:38] <memethuzla> ah, not version 5 ------ like I said, rating box lacks link to current dev page  ??? [20:38] <+Ironholds> yeah, that's unfortunate :(. I'll note that down as a suggestion for the future [20:39] <+Ironholds> I think the point is that the link gets traffic mostly from readers - around 1.5k hits a day - who are mostly looking for "what the hell is this?" and not dev info. [20:39] <jsalsman> this is going to be so much better than the 1/5 stars [20:39] <+Ironholds> but I've posted a link at the top of the "what the hell is this?" page, so it is now theoretically possible to get from the AFT box to the dev page [20:39] <+Ironholds> jsalsman: that's the aim :). Personally, as an editor, I'm really excited. [20:40] <+Ironholds> we can have feedback that users can actually respond to, that can't be gamed :) [20:40] <memethuzla> you are left having to look in two places unless you leave a not on 1st talk page & close it [20:40] <+Ironholds> memethuzla: what do you mean? [20:40] <jsalsman> you mean feedback being on a separate page from talk? [20:41] <+Ironholds> who was it earlier who said feedback should be a substituted element of talkpages, like wikiproject tags are? That's an interesting suggestion, and I want to be able to source it. memethuzla, was it you? [20:41] <memethuzla> ZSAorry, my response too slow. explicitly, the rate box need a link to the current device talk page. People at the old talk page need re-directing [20:42] <+Ironholds> ahh, gotcha [20:42] <Bensin> Ironholds: I made a comparison with the to-do-template. [20:42] <+Ironholds> Bensin: that's it! Noted down :) [20:42] <memethuzla> & no, not me (tags) [20:43] <+Ironholds> memethuzla: gotcha. well, it currently links to the current interface - that is, Version 4, the one that is deployed. It doesn't link to the Version 5 talkpage, just because like I said, it's mostly hit up by readers. But I will be letting all the commenters on the existing talkpage know that we've got a new talkspace, and I'm thinking of sticking an edit notice in. [20:43] <jsalsman> adding a link from the ratings box which appears on each page to an AFT5 (something like "preview changes to this tool"?) would be great [20:43] <memethuzla> good [20:43] <+Ironholds> So that every time you start a new section or whatnot on the talkpage it goes "are you here to bring up issues with the tool or suggest changes? If so go to [version 5 talkpage", something like that. Would that help? [20:44] <jsalsman> I think so. The sooner you get everyone involved the better [20:44] * Tiptoety|busy (~Tiptoety@wikipedia/Tiptoety) Quit (Quit: ChatZilla 0.9.87 [Firefox 3.6.23/20110920075126]) [20:44] <+Ironholds> jsalsman, memethuzla, I'll bring that idea (adding a link to the AFT) up with Howie. Like I said, that's mostly for readers, so it might not get done, but Fabrice was talking about trying to get readers to engage with the developers over this, so it's possibly a two bird, one stone situation. I'll run it by them. [20:45] <jsalsman> I wonder what "Details >>" and "Show all >>" would do on [20:46] <jsalsman> it makes sense to only show maybe the most recent 20 responses? [20:46] <jsalsman> unless they hit "Show all >>"? [20:46] <+Ironholds> I think that's the idea, yeah [20:46] <Bensin> Or perhaps top voted? If you let readers vote on comments. [20:46] <+Ironholds> if you see the bottom there's a "more posts" tab - I think that's to limit the number you have to look through at once unless you deliberately want to see the rest [20:47] <jsalsman> "(N) edits since this post" is great. I'm not sure whether "Moderation tools" is [20:48] <+Ironholds> Bensin: there is a plan to have "upvoting" and "downvoting" as well as simple hide/report abuse/whatever; I'm not sure whether it will be for editors only, or both editors and readers. It would make sense to have both. [20:48] <+Ironholds> while there could be a concern that it could be "gamed" (say, crazy white nationalist posts crazy white nationalist comment, is backed by his CWN bros and shoves it at the top of the Obama comments) the "hide" button probably makes it irrelevant. Any editor can just kill it. [20:48] <jsalsman> "Feature / List / Hide"? I guess those are the tools [20:48] <+Ironholds> jsalsman: indeed :). It's only a wireframe, so what "moderation tools" could be is still up in the air. What's your objection to it/ideas? [20:49] <jsalsman> I wonder if there will be comment hiding fights [20:49] <memethuzla> There's already gaming, I found some spectacular 5555s. Also, you can't vote 0 or -6 [20:49] <jsalsman> not sure what "List" is supposed to mean [20:49] <+Ironholds> memethuzla: indeed; gaming is one of the reasons we're moving away from simple stars [20:50] <+Ironholds> jsalsman: legit concerns. To be honest, I don't know what "List" means either :P. I'm writing this down and will drop you an email when I get some responses. [20:50] <jsalsman> memethuzla: presumably for numeric rankings, use of the median instead of the mean solves insincere outlier bias [20:50] <Bensin> Ironholds: IMHO voting for everyone. Editors will hide the comments that are not constructive criticism. [20:50] <+Ironholds> well, everyone games, to some degree. I always upvote *my* own articles [20:50] * jsalsman has no idea if AFT v4 uses median or mean aggregation but sure hopes it's median [20:51] <+Ironholds> although to be fair that's because I know they're really good :P [20:51] <memethuzla> again, in talk:Article_Feedback_Tool#Suggested_revision, I suggested a spoiler question to get the bias done with [20:51] <+Ironholds> Bensin: agreed. I'm going to bring it up. [20:51] <Bensin> Ironholds: HAHA!! [20:51] <+Ironholds> memethuzla: "are you a crazy russian nationalist?" [20:51] <memethuzla> not russian [20:51] <+Ironholds> no, no [20:52] <+Ironholds> I was joking that that could be the spoiler question :p [20:52] <+Ironholds> Bensin: so, if we're going to let everyone vote, we need some way of showing them the comments thread [20:52] <+Ironholds> my thought is that one of the options after sticking in comments should be "why not see what everyone else is saying?" or something. What do you think? [20:53] <jsalsman> maybe there could be a more detailed survey which attempts to determine whether the respondent has opinions about any issues on which there are general sanctions and if so puts their feedback in a separate "controversial"^W"independent thinker" column [20:54] <+Ironholds> maybe; that may be complicating it a bit much, though [20:54] <jsalsman> that would not go over well [20:54] <+Ironholds> I think generally if people are crazy, editors will spot it [20:54] <Bensin> Ironholds: The best would be to somehow let the reader first see the comments (top voted first) and _then_ decide if they want to "back" a current comment or post a new one. [20:54] <memethuzla> Ex.  ? ???????. ?? [20:55] <+Ironholds> we don't need a survey to determine that the man asking why our article on pharmacology doesn't include "BIG PHARMA KILLS PUPPIES" every 3 lines is two sandwiches short of a full lunch :P [20:55] * Jan_eissfeldt ( Quit (Quit: Leaving.) [20:55] <+Ironholds> Bensin: great idea! [20:55] <Bensin> Man, am I scoring big tonight or what? :-D [20:55] <jsalsman> it's probably good enough to just throw them all together with minimal moderation. I'd rather see the expected mess of rankings on Barack Obama's feedback than a scrubbed list that would necessarily waste editor time with hide/feature/list fights [20:55] <memethuzla> construct response cloud? [20:55] <+Ironholds> Bensin: seriously, you are [20:56] <+Ironholds> memethuzla: with words sized by number of iterations, one of those things? [20:56] <jsalsman> Bensin: yes that is good [20:56] <+Ironholds> like [20:56] <memethuzla> That' the thing. I gather computers can do that. [20:57] <memethuzla> exact [20:57] <+Ironholds> memethuzla: what can you see it being used for? [20:57] <jsalsman> wordclouds don't impress me. [20:57] <memethuzla> highlight specific observations about a page [20:58] <jsalsman> I want to say I really like Bensin's idea of showing top-ranked responses and allowing respondents to "+1" them [20:58] <+Ironholds> memethuzla: hmn. Possibly helpful - I'd like to think that if specific observations are good, they'd be upvoted though [20:58] <+Ironholds> jsalsman: I love it too :D. Definitely going on the list. [20:58] <memethuzla> both. cloud is just presentation [20:59] <jsalsman> I think that would be much better than hide/list/feature fights [20:59] <+Ironholds> memethuzla: I'll add it in, we'll see what they think. What's your username, so I have a way of contacting you? [20:59] <memethuzla> lists are so 19th C [20:59] <+Ironholds> hehe [20:59] <+Ironholds> okay, one moment - I need to write all this stuff down [21:00] <jsalsman> then the politician respondents could get two super-ranked comments about the subject, and editors could ignore those and scroll past to comments about the content [21:00] <memethuzla> memethuzla, memethuzla. But I may just have a little password problem after malware... [21:00] <YairRand> word clouds based on feedback sound like a bad idea to me. I can picture a group of people trying to see if they can get "nazi" up to 800px on certain blps. [21:00] <+Ironholds> jsalsman: I think there is an argument for hide features, but it's probably more for clearly abusive stuff [21:00] <jsalsman> agreed [21:00] <+Ironholds> YairRand: and (disclaimer: IANAT) on the tech side, calculating up to 3 million different wordclouds? recalculating whenever each one is made [21:00] <+Ironholds> ? [21:01] <+Ironholds> I mean, we get 40,000 comments a day at the moment [21:01] <+Ironholds> but I'll add it in anyway. [21:01] <YairRand> sounds completely impossible, but... [21:01] <jsalsman> the url blacklist is probably worth applying, but you said they wanted to apply the abuse filter. I can't remember if the AF does the blacklist these days [21:01] <memethuzla> don't need to recalculate evry time [21:02] <+Ironholds> jsalsman: they could do both - I'll find out. [21:02] <+Ironholds> okay, does anyone mind if I change the tack of this discussion slightly? [21:02] <+Ironholds> we've mainly talked about the voting and commenting system itself. Now. The existing plans, and the new ones - can you see things that need to be changed or added that would encourage readers to edit, not just to comment? [21:03] <+Ironholds> (one of the clear problems to me is that wikimarkup probably looks utterly alien to 80 percent of respondents, but that's not something I can fix directly) [21:03] <memethuzla> I had suggested "Are important relevant facts missing? (Please sign in, & share your knowledge.)" on main talk page [21:04] <YairRand> why sign in? [21:04] <memethuzla> to edit, to avoid idiots. [21:05] <YairRand> people often don't like to sign in. signing in isn't all that helpful. [21:05] <+Ironholds> memethuzla: I don't think that's necessarily a good idea [21:05] <+Ironholds> so at the moment around 2.7 percent of respondents manage to successfully edit [21:05] <+Ironholds> that's out of 17 percent who try [21:05] * fabriceflorin (~fabricefl@ Quit (Quit: fabriceflorin) [21:05] <+Ironholds> I think sticking another hoop up for them to jump through may hurt that number [21:05] <memethuzla> bad thoughts. Spammers dreams..... [21:06] <+Ironholds> memethuzla: yeah, but spammers can get through anyway :). That's what RecentChanges patrollers are for. [21:06] <memethuzla> A wysiwyg editor would change that overnight [21:06] <sumanah> "sumanah, do you know more about how that happened?" no, I don't. [21:07] <YairRand> and it all comes down to the visual editor [21:07] <+Ironholds> agreed, 100 percent. [21:07] <+Ironholds> the visual editor is coming. Just don't ask me when or in what form :0 [21:07] <+Ironholds> *:) [21:08] <+Ironholds> (the visual editor is also in the "not currently one of the things I'm working on" box. Most things are. I don't know much about it, as a result) [21:08] <sumanah> I know a little about the Visual Editor [21:08] <Bensin> I think AFT should be a fallback option for people who really just don't want to edit. I'm not sure why someone suggested AFT should be an on-ramp to editing. [21:09] <+Ironholds> sumanah: would you like to step in, then? [21:09] <+Ironholds> Bensin: I'll let sumanah step in for a bit, then answer that question [21:09] <+Ironholds> (if she's not busy) [21:09] <sumanah> Ironholds, Bensin, am reading scrollback to catch questions to answer [21:09] <+Ironholds> sumanah: gotcha [21:09] <Bensin> (A better interface for editing will probably do much more than a suggestion following a comment via AFT.) [21:10] <Bensin> (and that is coming, like you said.) [21:10] <memethuzla> but markup does impose a variety of biased IQ test [21:10] <+Ironholds> memethuzla: agreed. It's not even a particularly nice markup language ;p [21:10] <+Ironholds> there is actually some extension that allows for markdown, which is slightly nicer, but iirc it breaks everything too. [21:10] <sumanah> Brion Vibber, Gabriel Wicke, Inez Korczynski, Neil Kandalgaonkar, & Trevor Parscal are all working towards the new parser & visual editor [21:10] <sumanah> the latter depends on the former [21:10] <+Ironholds> Bensin: actually...did I ever tell you the Imperial College anecdote? [21:11] <memethuzla> Bensin: however an aim of rating is to engage new people [21:11] <Bensin> Ironholds: Don't think so. Please, go ahead! [21:11] <+Ironholds> so, Wikimedia UK did some events at Imperial College in London [21:11] <YairRand> so aft is still being developed and improved, and eventually it's probably going to be really useful, but approximately how much time are we going to have it and need it for, before the visual editor is launched? [21:11] <+Ironholds> and they ran into general members of the public (well, students, slightly less washed than the general public) and asked them if they wanted to edit, why they hadn't, etc, etc. [21:12] <+Ironholds> YairRand: if you let me finish my story, you'll see that both an on-ramp and a visual editor are necessary [21:12] <+Ironholds> YairRand/Bensin: can you guess what the number one response to "why don't you edit" was? [21:12] <+Ironholds> and note that these are students at one of the most prestigious science unis in the UK [21:12] <YairRand> "You can edit?"? [21:12] <+Ironholds> YairRand: precisely! [21:12] <Bensin> "Can I really?" "Don't dare" are usual answers. [21:12] <memethuzla> to drunk? [21:12] <memethuzla> too drunk [21:12] <+Ironholds> it was " mean you can EDIT? Like, anyone? I thought they had, y'know, special people..." [21:13] <sumanah> I have run into the very same thing [21:13] <sumanah> many many times [21:13] <+Ironholds> despite what we might think, a lot of readers genuinely don't know they can contribute. That's one of the reasons an on-ramp is so great, *if* we can get it right. [21:13] <+Ironholds> and like I said, this isn't Aunty Ethel from No. 54 who still has to unplug her phone to use the internet. [21:13] <memethuzla> "special"? [21:13] <+Ironholds> these are highly intelligent engineering, physics, chemistry and comp sci students [21:13] <jsalsman> to be fair, the logo which said "The encyclopedia anyone can edit" now says "The free encyclopedia" [21:14] <+Ironholds> and they didn't know! [21:14] <+Ironholds> jsalsman: this is true [21:14] <memethuzla> tip-ex, anyone? [21:14] <+Ironholds> but yeah; a lot of readers genuinely don't know they can contribute. This is why the AFT and an on-ramp is so great, because it actively asks them "did you know you can edit? Would you like to?" [21:14] * fabriceflorin (~fabricefl@ has joined #wikimedia-office [21:14] <sumanah> memethuzla: please clarify? [21:14] <YairRand> "Wikipedia, the free encyclopedia that anyone can edit. Including you. Seriously. No, we're not nuts. Well, maybe a little." [21:15] <+Ironholds> YairRand: "okay, we're completely nuts" [21:15] <+Ironholds> "no, seriously, we have one guy? ALWAYS dresses in orange." [21:15] <jsalsman> can we try an A/B test with the old logo and see what it does to IP edit rate? [21:15] <memethuzla> the tip-ex is so difficult to get off the screen (Essex girl joke) [21:15] <+Ironholds> jsalsman: I can pass it on, but it may be a bit out there. No promises. [21:15] <jsalsman> or at least the old logo caption [21:15] <Bensin> +1 on jsalsman's suggestion. [21:15] <sumanah> so, any questions about the Visual Editor? I can help answer or point people to answer-y resources [21:16] <sumanah> unless that is offtopic, Ironholds, fabriceflorin [21:16] <+Ironholds> sumanah: I think this was generally about the AFT, and we sort of segued. I have to leave and wrap this up, but you are welcome to chat about the visual editor, of course. [21:16] <sumanah> Sure. [21:16] <Bensin> sumanah: Aren't there more than one visual editor? Do you have a link to the one you work on? [21:17] <sumanah> So, when I refer to The Visual Editor, I'm talking about the editing interface that the people I mentioned are working on, at or with the Wikimedia Foundation. [21:17] <sumanah> [21:17] <sumanah> as part of [21:18] <+Ironholds> okay, I think we have to close this down - I need to go off and type all these ideas up. Thank you so *so* much for all the suggestions. You can catch me at if you have any further suggestions, since my Foundation email address isn't working yet, although I'd encourage you to comment on [21:18] <YairRand> is the real-time collaboration thing part of the visual editor? [21:18] * gwicke (~gabriel@ has joined #wikimedia-office [21:18] <sumanah> Bensin: As volunteer development coordinator, I talk to Neil, Inez, Gabriel, Brion, & Trevor and get updates from them for [21:18] <sumanah> which you can subscribe to [21:18] <Bensin> Ironholds: Thank you for setting this up and for listening. [21:18] <+Ironholds> I can't promise you all that the community input process for the AFT will be perfect, or that you'll get everything you want. In fact, some bits will probably suck, and we *will* screw up, because we're human. [21:18] <+Ironholds> but I can promise that we're going to do it *better*, even if we don't do it perfectly. [21:19] <jsalsman> \o/ [21:19] <+Ironholds> Bensin, YairRand, jsalsman, memethuzla, thank you for all your helpful suggestions (and the essex girl joke ;p). I'll drop you all notes with details on how you can further engage, and responses to the suggestions you made. [21:19] <sumanah> YairRand: so is related to the visual editor work [21:19] <jsalsman> thank you, Ironholds [21:19] <+Ironholds> in the meantime, like I said, if you find anything email me or drop stuff on that talkpage. Take care all :) Session Close: Thu Oct 27 21:20:00 2011