User talk:Jdforrester (WMF)

Hello! Please feel free to leave me a message here. -- James F.

Just a note

Hi James. I just wanted to take a moment to leave a note here for you, and please feel free to pass this on to your team, since it applies to them as well. So here goes: Thank you for working on VisualEditor. Thank you for putting in what I assume are incredible amounts of man- and woman-hours, and for taking a real step toward deploying something that's been essentially vapourware for years. Whether this deployment was perfect or not, you guys have put in the time to try to get a usable visual editor up and running, and it's unfair of the community to treat you as if you spent all that time cackling evilly over how to break things as badly as possible. Disagreement is one thing, but calling for heads and spewing insults is another, and on behalf of what I expect is most of the community, I'd like to apologize for the gratuitous nastiness you've been subjected to by some users. Please don't be too discouraged by or afraid of the loudest angry voices; for every one person shouting incoherently, there are many who just want to help move past what went wrong here and toward a workable deployment of VE, and we need you guys to work alongside us without fearing us (or resenting us) to accomplish that. A fluffernutter is a sandwich! (talk) 17:25, 24 September 2013 (UTC)

@Fluffernutter: Thank you very much. Jdforrester (WMF) (talk) 19:08, 24 September 2013 (UTC)

Fluffernutter, feel free to give your own opinion, but please don't try to speak "on behalf of what I expect is most of the community". We have here a person whose job description is "My job is to help make sure the VisualEditor team understands what the community wants and needs, is focussed on the things that matter, and is engaging with and understood by the community." It is clear that he failed at that job quite spectularly the last few months, and comments like (among many others) "we cannot justify continuing to exhaust staff time (read: donors' funds) to this issue" only stress the gap between his job description and his actual actions. Someone who beliefs donor's funds are more important than the opinion of the largest Wikipedia version (shared apparently by the second largest one, the German Wikipedia) and gives that as the main or only reason to finally implement the solution this community wanted (and had implemented already anyway without the WMF), is not even trying to be "understood by the community", and clearly doesn't care about "what the community wants and needs". Donors give money because they like Wikipedia and want it to continue; they don't give money to support devs against the wishes of the editors. That you seem to value money and donors more than volunteer time and efforts is not compatible with your function or with the fundamental principles behind Wikipedia and Wikimedia. Fram (talk) 07:47, 25 September 2013 (UTC)

Fram, I was apologizing because I think it's been highly inappropriate of some community members to call staffers (or anyone, come to think of it) names or treat them as though the VE debacle is something they did to troll us rather than something they spent genuine time and effort on because they thought it would be useful. I'm not apologizing for there being community discussion or controversy over the WMF's actions, and I don't think anyone on the WMF end would expect apologies for that; I'm just apologizing for the fact that some people veered away from discussing the issues to instead personally attack staffers and their motivations. A fluffernutter is a sandwich! (talk) 14:43, 25 September 2013 (UTC)

Sentences like "it's unfair of the community to treat you [...]" gave a distinctly different impression. Or things like "for every one person shouting incoherently, there are many who just want to help move past what went wrong here and toward a workable deployment of VE"; it gives an "either you are shouting incoherently, or you are willing to work with the WMF" division, while I believe that most critics don't belong to either camp at the moment (I mean here that many are not willing to work with the WMF if things don't change drastically, not that they don't want to work with the WMF ever). I have no problem with you indicating that the people simply insulting WMF staffers for the sake of it are a tiny minority though, and that they shouldn't keep WMF and others away from discussions. Fram (talk) 14:50, 25 September 2013 (UTC)

I agree with Fluff. I've just read the thread on AN; it was by the same user who took down Anobody, so I suppose one shouldn't be too surprised they were able to powerplay the Foundation. Still rather shocking though. VE was a really promising start. Thanks for all the work you and your team put into it. I hope you'll be able to return it to us soon. FeydHuxtable (talk) 21:02, 25 September 2013 (UTC)

FeydHuxtable, it's still being worked on and it is available as an option for you to edit and report bugs on/make feature requests on as well. See WP:WMF for more details about reporting bugs/requesting feature improvements. See Special:Preferences#mw-prefsection-editing to turn it on. That link is at the top of WP:VE/F. Best. Biosthmors (talk) pls notify me (i.e. {{U}}) while signing a reply, thx 21:07, 25 September 2013 (UTC)

Feyd, please don't rewrite history. Anyone who was around at the time will know that the only person who "took down" ANobody was ANobody himself. Persistently disrupting Wikipedia, refusing to answer community concerns about his editing (remember the RFC/U?) and then socking is generally going to get you blocked, whoever you are. Black Kite (talk) 18:16, 26 September 2013 (UTC)

Edit summaries in VE

You said in Wikipedia talk:VisualEditor/Default State RFC#Wikimedia_response, as an example of where VE showed an improvement over wikitext editing:

"One thing we have noticed, which we feel is beneficial, is that users are 6 times more likely to use edit summaries than with the markup editor – though earlier in VisualEditor's development it was more than 10 times more likely, and we're investigating what led to the shift."

You have been asked repeatedly, at different venues, to explain this statement and to back it up with some evidence, as it seems highly unlikely to be true. As far as I have seen, you haven't given any response to this. Could you point me to your response or provide one here please? Fram (talk) 07:09, 27 September 2013 (UTC)

@Fram: Hi! Sorry, not seen the requests for this data. Will go and dig it out again (this was noted in July, but we've been a bit busy since then). Off the top of my head (cn ;-)) it was roughly a 7% -> 85% change if you include section links, and 7% -> 45% when you take out section edit links' auto-summaries where nothing other than the auto-summary has been added. This fell to ~25% when we started describing the edit summary box as such, rather than just leaving it unlabelled, hence my particular interest (i.e., that removing items that the community expects for totally reasonable causes to be in the interface can actually drive up desired behaviours, not down). Jdforrester (WMF) (talk) 16:15, 27 September 2013 (UTC)

Thanks. Seems like wrong data in your statistics though, every test I have seen (samples made by different people) had the edit summary rate more like at least 20% without VE and excluding all automated and section ones, and a bit more with VE. That something is wrong with those data is clear when you notice that for wikitext, you get the same 7% no matter if you include automatic ones or not. E.g. from the latest 100 edits in article space, only 20 have no edit summary whatsoever, so 80% has some edit summary, either automated or not. The chance that once again, my sample has found a statistical anomaly is minimal... Next time, when you include such figures in the justification for something the WMF wants, at least include some link to where you got those figures from. And perhaps also include numbers that point in the other direction, like the one from the WMF that shows that for new editors, 80% used wikitext and only 20% VE (before it became opt-in of course). Fram (talk) 15:15, 28 September 2013 (UTC)

@Jdforrester (WMF): Any link to the data you used to get these numbers? Fram (talk) 08:12, 8 October 2013 (UTC)

@Fram: No. I'll ask a colleague in Analytics if we can re-run those numbers for you, but it's not a priority, sorry. Jdforrester (WMF) (talk) 19:12, 8 October 2013 (UTC)

OK. I don't get why numbers which are used to justify something controversial like VE (even as an example only) are not kept somewhere. It gives the impression that they are invented to impress us (even if the save mechanism of VE could be included in wikitext editing anyway, if that was the major improvement the WMF saw), which rather backfired since they seem to be wrong. If the WMF wants us to trust them again, they will have to do better than this. Fram (talk) 07:01, 9 October 2013 (UTC)

SUL Finalisation

Hi James. Hope all is well with you. Just checking in regarding SUL finalisation. Any idea when that's now likely to happen? Best, Will - WJBscribe (talk) 16:35, 27 September 2013 (UTC)

@WJBscribe: Hey WJB; that's now mostly @DGarry (WMF):'s role, though I'll be helping him. Hopefully we'll be able to make some progress soon. Jdforrester (WMF) (talk) 23:53, 27 September 2013 (UTC)

Hi WJBscribe. To expand on this a bit, the finalisation is is on my list things to do. It's quite high priority as it's been hanging around in limbo for a while. That said, Howie, Rob and I decided that for the immediate future (the next few weeks or so) it would be best if I focussed on a single task (mw:Auth systems) so I could build up some experience. I'll try to get to this as soon as I can. --Dan Garry, Wikimedia Foundation (talk) 09:34, 28 September 2013 (UTC)

VisualEditor / Status

I notice that you have updated [1] the VisualEDitor / Status page at MediaWiki. Any reason that you haven't changed the V19 status to more accurately reflect what actually works and what doesn't in it? I have left feedback at the two pages where you requested it (WP:VPT and WP:VEF), pinged you, and tried unsucesfully to change the status page myself, only to be met with extreme hostility and a very un-wikilike environment. Any reason why the status shouldn't reflect the current knowledge about what works and what doesn't, certainly when the version still needs to be rolled out to most wikis? Fram (talk) 08:23, 1 October 2013 (UTC)

(replied there Jdforrester (WMF) (talk) 01:32, 5 October 2013 (UTC))

Office hours

I have to say, I'm glad that I didn't join the office hours, reading the logs seems to indicate that they were a total waste of time. Nice to see that Spanish editors have the same problems and get brushed off just as easily though.

You claimed in the office huors log that "But in 3/6/12 months it will not be true, and it will confuse people. VisualEditor is already proving to be a success for new users, despite the issues." What do you consider "proving to be a success for new users", apart from anecdotical evidence? WMF figures indicated that (before the opt-in was enabled) newly registered editors used VE for 20% of the edits, and Wikitext for 80%. No significant change could be detected in these figures over the months, there was no increase in VE use. I have anecdotical evidence (no means to collect this more thoroughly) indicating that people tried to use VE, failed, and then simply stopped editing. Are such things monitored and taken into account when you claim that is "proving to be a success"?

Just like with the edit summary issue above, please provide some good evidence for your claims or don't make them please.

In the same log, you claim "Also, the numbers of accidental mis-formats (having to fix your syntax after saving) is zero with VisualEditor, obviously." Um, what? What do you mean when you make such an outlandish claim? Even today, basic edits by people who regularly use VE for simple things create malformed pages, like here (mentioned on WP:VEF). If this is not what you mean, then please give examples of what you mean with "fix your syntax". Removing empty section headers? Needs to be done in VE. Saving internal links as external links and vice versa? Gets done with VE all the time. Anything else? Fram (talk) 13:28, 9 October 2013 (UTC)

I believe that "accidental mis-formats" refers to things like [unpaired brackets]] in a failed effort to make a link. BracketBot sends out a couple hundred messages every day about that one problem (a number of messages, BTW, that exceeds the number of nowiki tags produced by VisualEditor each day before it was changed to opt-out). The classic wikitext editor is essentially "designed to fail" whenever the user makes a typo with brackets, or forgets the slash in the second ref tag (thus putting all the text between that typo and the next ref tag [or the entire rest of the page, if it's the last one] into a footnote), or otherwise doesn't get every single character of formatting typed correctly.

It's also worth remembering that different wikis have different responses. I checked every diff made in VisualEditor at ja.wp for weeks during the summer, and I never once saw the nowiki problem: Japanese editors just didn't type [[bracket]]s inside VisualEditor. At the Spanish Wikipedia, VisualEditor is currently (last ten days or so) being used by slightly less than half of the new users (accounts registered since their rollout date on July 25th). When you keep in mind that about a quarter of users can't even see VisualEditor (because they're running IE or have Javascript disabled), I think it is not unreasonable to interpret this as a clear majority use of those editors who are actually able to make a choice for that project. What happens at the English Wikipedia, with the oldest editor base and by far the most technically complicated pages, is not what happens everywhere. Whatamidoing (WMF) (talk) 15:58, 9 October 2013 (UTC)

[ec] User:Whatamidoing (WMF) is indeed correct - what I had written:

@Fram: Sorry for my inclarity. It is fundamentally impossible in VisualEditor to make the very common wikitext mistakes like unbalanced bold/italics (e.g., ''foo'''), broken template calls (e.g., {{foo}), broken references (e.g., <ref>''foo''</ref>), broken image invocations (e.g., [[File:Foo.jpg|Caption|thumb|right]]) broken categories (e.g., [[Category|Rivers in Egypt]]</code>), mis-matched reference names (''e.g.'', <code><nowiki><ref name="Foo">Example</ref><ref name="Fo" />), mis-typed template parameter invocations (e.g., {{infobox|nam=John Smith}}) and dozens of other hugely common mistakes that catch out new and experienced editors alike, and which again and again and again have been demonstrated to drive new contributors away. Jdforrester (WMF) (talk) 16:06, 9 October 2013 (UTC)

But on the other hand, it is much more common now to see half-linked words, empty section headers, text converted into a section header, and so on. Wrong edits like this one are easier to make in VE than in Wikitext. Having prevented a number of errors doesn't mean that "the numbers of accidental mis-formats (having to fix your syntax after saving) is zero with VisualEditor, obviously.", and such claims can only backfire. Whatamidoing, note that the problems / backlash isn't only here, but also on the German and Dutch wikis. It seems as if it not some purely local problem. I have heard that both the Spanish and Danish wikis also were discussing changing to opt-in, but I have no idea what the current status on this is. Are any figures available on the percentage of users that opens a page in edit mode (VE or wikitext), but don't save their edit? Is this higher or lower in VE? Do you have a number of users that made an edit in VE, and then made a consecutive edit on the same page in the Wikitext editor (indicating that they probably couldn't do whatever they wanted in VE)? Fram (talk) 08:25, 10 October 2013 (UTC)

They ran those numbers at the start of July, and discovered that the system had some problems (like more people saved an edit in VisualEditor than opened it, or something similarly implausible). I don't know what happened since then. I do know that they ought to exclude me from any such report, because I open VisualEditor to verify bug reports multiple times a day, and I save very few of those edits. Whatamidoing (WMF) (talk) 17:06, 10 October 2013 (UTC)

Thanks. While I would obviously prefer to have such figures, I appreciate a honest answer that an attempt to get them failed, and so you don't have them. Can't be helped :-( Fram (talk) 20:06, 10 October 2013 (UTC)

Note that I have no objections against implementing the few improvements VE made into the wikitext editor. E.g. the template editor, with Template Data, could be a good improvement for Wikitext editing (making the current citation drop down menu more general in fact). On the other hand, the file treatment, which may have eliminated some errors but at the same time eliminated nearly all functionality, is not an improvement. I see no reason though why you couldn't have a dropdown "files and templates", where the options for a file are treated like the parameters for a template, and you can indicate "thumb", "left" or "right" and so on. Fram (talk) 08:58, 10 October 2013 (UTC)

I can't comment on the Japanese wikipedia, but looking at the Spanish, where the "nowiki" tag doesn't seem to be activated, the most recent VE edit is this, adding incorrect nowikis. No idea why this doesn't happen in Japan, but it clearly isn't an enwiki-only problem. (also e.g. here. And this may not be a syntax error, but it's hardly good code either (and rather rare in Wikitext editing). I also note a very high vandalism / vandalism reversion rate in these edits, but I haven't compared this to wikitext editing so I can't claim that this is VE related. Fram (talk) 09:37, 10 October 2013 (UTC)

@Fram: I'm immensely puzzled by this - VisualEditor has not remotely "eliminated nearly all functionality" from media files. We have repeatedly and clearly said over the last nine months that the rest of the functionality (all of it) is coming, but is less of a priority. If you're going to mis-represent that, how can we have a proper conversation about the matter on which we should focus and when? Jdforrester (WMF) (talk) 18:13, 11 October 2013 (UTC)

Pot, kettle, black. In my opinion, yes, it has "eliminated nearly all functionality". What can I do? I can add a file, I can resize it, and I can give it a caption. I can not chose the position, I can not edit any of the parameters directly (alt text? Size? see Wikipedia:Extended image syntax), I can not change the file (replace file 1 with file 2)... And I have absolutely no idea how I am supposed to add or modify a gallery in VE (which is "also" file functionality!). And the new fucntionality (resizing, moving) works so poorly that it is hardly worth it. The only improvement I see so far is the file choice display when you add a new file. Please, next time, try to remember that a difference of opinion is something completely different than misrepresenting things. For me, nearly all functionality wrt files is missing. For you, it isn't. Fine. If you can't live with someone having a different and less favourable opinion, then yes, it will be hard to have a proper conversation, and it will be hard for you to do your WMF job properly. But don't blame me for your problems. I'm one of the few editors on enwiki still testing VE and commenting on it. Perhaps try to encourage others to join, instead of alienating the few that are left? Fram (talk) 07:22, 14 October 2013 (UTC)

"Eliminated all functionality" makes it sound like VisualEditor previously supported all these things but is now actively removing these elements (e.g., automatically deleting the |left position setting or the alt text on an image), not merely that the current version doesn't happen to let you edit that part of the page. Whatamidoing (WMF) (talk) 18:39, 14 October 2013 (UTC)

Well, I thought it wsa obvious that I meant "eliminated" compared to what we had (i.e. wikitext editing), not that VE used to be better somehow. But if this was the cause of the confusion and rather vehement reply, then thanks for clearing this up. Fram (talk) 09:37, 15 October 2013 (UTC)

An example of things that are "fundamentally impossible in VE" and happen zero times: this edit empties a named ref, causing the "Cite error: The named reference prisa was invoked but never defined" error. No idea how the editor did it, I haven't tried to reproduce it, but it shows that VE also enables (and sometimes produces) such edits. Making categorical claims is rarely a good idea. Fram (talk) 08:19, 21 October 2013 (UTC)

It appears that you can reproduce this if you select the ref, open the ref dialog, blank the contents, and apply your changes. This can be reproduced both when the named ref is reused and when it's not, as well as when the ref is added in VisualEditor and blanked during the same editing session (not saved, so no diff for you, but it looks identical).

I'm not sure what the expected behavior should be: to have the whole ref removed, even though the user didn't delete the whole thing? to do what the user actually did, which is blank the contents? to produce some sort of warning? Whatamidoing (WMF) (talk) 20:58, 21 October 2013 (UTC)