User talk:LilaTretikov (WMF)/Archive 11

Upcoming 5 millionth article milestone on English Wikipedia

You are invited to participate in the discussion at https://en.wikipedia.org/wiki/Wikipedia:Village pump (miscellaneous)#5 millionth article celebration: community press release, video, open letter, and site banner. --Pine^✉ 07:03, 14 September 2015 (UTC)

Thank you, Pine! LilaTretikov (WMF) (talk) 18:22, 14 September 2015 (UTC)

The Wikipedia as a whole with all languages included contains close to 10 million articles. Whether it is above or below probably depends on the definition of "article" and whether or not bot generated articles are included. I'd be pleased, if the WMF would officially celebrate this global feat -- globally :-).---<(kmk)>- (talk) 21:14, 14 September 2015 (UTC)

Samples of some of the Flow issues

(copied from a section earlier on this page)
Alsee can you please provide URL/phabricator to reproduce. If what you are saying is true, I am not sure how people are using it at all.  LilaTretikov (WMF) (talk) 21:27, 16 September 2015 (UTC)

• Undo mangles content, the Phab link was pasted earlier. https://phabricator.wikimedia.org/T112338
• Copy-paste mangles content, I don't know aboout Phab, but Quiddy and I have a pile of posts examining it here.
• Switching </> from wiki to visual destroys content. This appears in in the above thread. I paste a wikitext table into wiki mode, switch to visual, the entire content gets mangled down this empty version:

{| style="margin: 1ex auto 1ex auto" class="wikitable"|}

• Switching </> from visual to wiki and back to visual, mangles what you had in the editor. To reproduce go to Visual mode. Type foo enter enter - enter enter bar. Note you have blank lines above and below the hyphen. Click </>. Click </>. Note that the blank lines are gone.
• No oversight: I asked Keegan to oversight a post that inadvertently captured my watchlist token. Oversight had no effect. Keegan then tried to delete it. No effect. Eventually Roan Kattouw went into the database and manually killed the pointer to the content. The result was that using history and following the link to that post completely bombed.... listing a Fatal Exception code and offering a link to Main Page.
• Losing access to Hide and other menu options: While working with Quiddy on the copy paste issue, in desperation I tried CTRL-A CTRL-C on an entire rendered page containing the test table. (This captured a link containing my watchlist token.) I pasted it into Flow to see if the table on that page would render. I guess the content was so "big" that it reached up and covered the menu for Hide and other options. So no Hide, no Oversight, no Delete. My private watchlist token was stuck on display. The watchlist token is a minor privacy issue, but it could just as well have been more serious oversight material.
• Mixing top posting and bottom posting turns complicated discussions into spaghetti. I picked a modest size Talk page discussion (~40 posts). I replicated it in Flow. You can see it here. If I ran the test again I would change where some of the replies were attached, but the general issue was clear. Posts at the bottom discuss or decide issue, then someone replies to a person earlier in the thread and gets top posted. If they explicitly or implicitly respond in light of the older posts, their top-post is incomprehensibly out of context.
• Not a "bug" per se, but I used that Talk page simulation to calculate Flow's information density.
  • Assuming Wikitext = 100%
  • Flow default mode = 41%
  • Flow wide mode = 48%. (Note: Wide mode loses all of the headers or other static content on the page. )

Awful.

Yeah yeah, I read the FAQ:Why does it look like Facebook. I'm telling you, it's wrong. I've seen people citing this as a comical example of how out of touch the WMF and Flow are. Go to your programmers, tell them you're giving them a new development environment. Tell them that the new programming environment only displays 48% as much code on the screen at a time. Then tell them that the decision is based on "elaborate [Facebook] user research and A/B testing to find what works best". They won't be amused.

• This issue is needed to view the next next one. In Visual type star space foo enter bar enter. This gives:

• Foo
• Bar

Then try to put some blank lines in between Foo and Bar. You'll get this:

• Foo
• (empty bullet point)
• (empty bullet point)
• (empty bullet point)
• (empty bullet point)
• Bar

Visual did not want to let me remove the empty bulletpoints. Deleting doesn't work because hitting enter just brings them back. I had a whole battle with the editor, trying to get rid of them. Anyway, copy it without saving. Start a new post and paste it in. Note that you now have blank lines between the bullet points. (WTF?) Save, the page renders with the blank lines gone.(WTF?) Click to edit the post. The blank lines are back.(WTF?) This is what I referred to as "low-grade sort of mangling". But I don't know how deep that issue goes. When I tried more copy-paste tests I got such a rapid chaotic set of results that I can't file a bug report any more specific than copy-paste is completely broken. Bullet points would vanish, indentations would change, and other random mangling. I have no idea how much mysteriously changes on save or re-edit.

• This test demonstrates Flow failure trying to delete a single disruptive comment from the middle of a discussion. It leaves a link "This comment was hidden by Alsee (history)" in the middle of the discussion, which is a giant neon sign saying "CLICK ME! CLICK ME!".
• This test demonstrates that I can't effectively delete bulk disruption/spam.
• Flow won't let me refactor a discussion. In Wikitext I can trivially move a single a single post, trivially move a block of posts, put a show/hide template to collapse a block of 10 offtopic posts, I can run down a list of replies and change indentation.
• This post shows the two incompatible :*# semantics, and as a bonus it shows the crazy range of results you get when you try to copy from visual/wikitext/rendered and paste to visual/wikitext. Colon has different indent sizes in the two semantics, bullet and list have different character-inputs in the two semantics, and only one of the two semantics supports nesting. And Flow tries to embed both semantics into the same post at the same time.

I don't recall what other random issues I ran into. And as a reminder - I Don't Use Flow. The only time I've used Flow is a handful of tests, and a few times I had to use Flow to post problems on DannyH's page. Every time I touch Flow I run into new Critical Failures and assorted smaller issues. How on earth am I the first person to discover copy-paste and UNDO are catastrophically broken?? How are the Devs unaware that switching </> mode can mangle content? How did no one check whether the threading model works? The developers were told to build a chatboard - and they succeeded. Then kludges were piled on. Flow started on a mistaken theory, ran off in a mistaken direction, and the kludges trying to make it work are popping at the seams. Alsee (talk) 02:03, 17 September 2015 (UTC)

Today I decided to test the Mark-As-Resolved option. In a matter of moments I ran into:

An error has occurred.
Return to Main Page.
[a3b876d0] 2015-09-19 21:27:46: Fatal exception of type "Flow\Exception\FlowException"

I don't know yet if this is a generic history error, or if it was specifically related to my Mark-As-Resolved test edits. Here is the Flow thread where I posted the issue, and here is the link that explodes in a Flow exception.

Every. Time. I. Test. Flow. It. Explodes. Alsee (talk) 21:37, 19 September 2015 (UTC)

A bold thought experiment

Lila, as an exercise in lateral thinking when people say that we cannot do something, I try to consider what would happen if we did do it. In that spirit I would like to pose a little experiment. You wrote earlier that "not all messages can get an answer from staff". Suppose that the WMF was an organisation in which every message could get an answer -- what would it look like? It would probably have a few clear lines of communication, well-defined avenues and well-designed portals, well-trained and effective moderators managing discussion and significant and innovative technical infrastructure. It would probably also have a much higher degree of community engagement with a more positive and enthused community, and it would be much more efficient and much more effective. It would not get itself into the sort of situations I mention above under "Us and them", and all the other unhappiness that you have spent a lot of time here on this page coping with. In fact, it would be an organisation in which you had a lot more time to think and plan. Surely that is worth aspiring to, and spending at least a little effort trying to achieve? Would you like to set up a little cross-community group to come up with a description of what such a WMF might look like and how we might get there? Rogol Domedonfors (talk) 21:14, 31 August 2015 (UTC)

Greetings Rogol. I don't mean to intrude but I noticed that most of the comments on this page are from only a small handful of people - you being one of the more prominent ones. It is certainly your prerogative to ask these questions and have a cordial, civil discussion, but there is something to be said about monopolizing it as well. Most questions here seem valuable and I hope we don't lose them in the process. I don't know Lila's schedule but I assume she is quite busy, more and more comments in shorter spans of time isn't likely to get her to engage any faster, it will only make the task more time consuming till it seems like a mountain to engage and research and answer these points, which would be a net loss for everyone involved. I hope you don't take any umbrage to my observation, after all I'm speaking from experience of doing the same ;). You might also want to consider approaching the concerned staff members or the board members for particular issues - they would have more hands-on knowledge and might be able to offer a quicker response time. Kind regards. Theo10011 (talk) 22:15, 1 September 2015 (UTC)

Thank you for that. If Lila wishes not to receive constructive comments from me or anyone else I am sure she will not hesitate to say so. It is regrettable that her staff are often too busy to interact with ordinary volunteers (WMF)&diff=prev&oldid=12643209, as indeed are the Board members [1]. You will have noticed that this discussion is on that very topic. Rogol Domedonfors (talk) 21:15, 4 September 2015 (UTC)

User:Rogol Domedonfors I am a board member and am following this conversation. The last couple of weeks I have had really poor internet access. I do agree that greater clarity on the relationship between the community and WMF is needed.

IMO the community needs to approach the WMF when they have clear consensus. And when the community does approach the foundation with clear consensus the foundation need to address it. We must made sure no "one ordinary volunteer" gets to speak for all volunteers as often volunteers have contradictory opinions. Doc James (talk · contribs · email) 10:18, 22 September 2015 (UTC)

@Doc James: But when the community does approach the WMF, even with the outcome of an RfC, the WMF doesn't listen. See for example VE and MediaViewer on enWP. Also at mw:User talk:DannyH (WMF), Quiddity said, and I quote, An RfC isn't needed, because we already know that many editors want as little change as possible. The clear implication is that the WMF believes that any community consensus procedure, of which RFCs are king, will (in their opinion) irrationally reject any attempt at (in their opinion) progress and that they are therefore not worth listening to. This was the same reasoning given for rejecting the enWP MediaViewer RfC. So what's the point of arriving at a clear consensus when it won't be listened to? BethNaught (talk) 13:12, 22 September 2015 (UTC)

Yes if the WMF rejects a large community consensus without justification I agree we have a problem.

DannyH also says "clearly establish how many of... the broadly desired improvements and expansions... are possible using a plain wikitext database"

I take this to mean that if it is possible to make individual sections of talk pages watchable than they will do this. If it is possible to make notification work between WPs they will do this also. Making individual sections watchable however they say is hard. Doc James (talk · contribs · email) 14:15, 22 September 2015 (UTC)

@Doc James: I don't think I have received a clearer or more patronising exposition of the stance of the WMF senior management and the Board than this. As I understand your position is this. A community consensus, whatever that is, may receive a response, but "one ordinary volunteer" who dares to meddle in matters of high strategy should expect to be ignored, which is what you think they deserve: it goes without saying that they can have nothing useful to contribute. Individual volunteers who ask staff or Board members to deliver on or explain progress on commitments made in public to the community are nothing but nuisances who deserve to be and will be ignored at best and insulted at worst. In short, ordinary volunteers should shut up and go away. I seem to recall that you stood for election on a platform of "prioritizing a better relationship between the editor communities and the WMF". I should be interested to hear how you think your patronising dismissal of the "ordinary volunteer" will help to achieve what you claim to desire: since you have chosen Lila's talk page to publish your views, I should be interested to hear whether she agrees with you. Rogol Domedonfors (talk) 18:26, 22 September 2015 (UTC)

Your interpretation of what I wrote is not accurate and does not reflect my position.

I would not call requesting consensus/support for a new change a patronizing dismissal of an individual editor. If a single editor puts forwards the idea to abolish abrcom they should produce substantial support before the idea is investigated or acted upon by the foundation. This is what we require within the projects before new major changes are cared out.

When a clear consensus is brought forwards such as that superprotect should not be used to roll out software, this does need a clear response and engagement from the WMF.

Yes of course the ideas of individuals editors matter. When the request is to provide details of previous promises these requests should be addressed. Doc James (talk · contribs · email) 22:26, 22 September 2015 (UTC)

I am glad to hear that you believe that the ideas of individual editors matter. You will have gathered from this discussion that your belief is not shared by some of the WMF senior management or your Board colleagues. This discussion was a suggestion to Lila for a thought experiment about how that might be improved. For some reason it got diverted by others into a somewhat unhelpful debate about whether certain individuals who troubled to propose ideas to WMF leadership were being presumptuous. To get the conversation back on track -- do you have any comments on the original suggestion? Rogol Domedonfors (talk) 06:12, 23 September 2015 (UTC)

I have mulled around the proposal of all staff having a certain percentage of their time to engage with the community per here. Or at least someone within each department having direct community involvement. I think that not only would greater communication be useful but clear details regarding what forms of decision making holds the greatest authority. The foundations of our organization include both that of a "do-ocracy" and that of a "direct democracy" Doc James (talk · contribs · email) 18:14, 23 September 2015 (UTC)

ECJ opinion

Lila, the European Court of Justice has issued an opinion on the question of privacy law applying to corporations that share personal data between EU and US sites [2]. Do you expect to have to make any changes to WMF Privacy policy as a result? Rogol Domedonfors (talk) 21:27, 23 September 2015 (UTC)

PETA vs WMF

Lila, I see that PETA are proposing to ask a Federal court in San Francisco to grant copyright in the famous so-called "monkey selfie" to the monkey and to have themselves declared the monkey's trustee to receive her royalties [3]. Given that this photograph is deemed by Commons to be in the public domain, there is a clear threat to the project's interests if that case should succeed. Will the WMF be filing a brief in this matter to support the Commons position? Rogol Domedonfors (talk) 21:17, 22 September 2015 (UTC)

It would be crazy if a US activist group was able to claim ownership of all pictures taken by non human animals. I find it hard to believe that this is anything put a publicity effort. Doc James (talk · contribs · email) 16:21, 23 September 2015 (UTC)

Publicity stunt, maybe, but presumably they'd love to establish legal precedent for granting person-like rights to animals. (Looking at the rights SCOTUS gave corporations, why not?) --Pi zero (talk) 16:51, 23 September 2015 (UTC)

A non-profit spending undisclosed amounts of donors' money on a meritless legal action for the publicity -- surely not! However, the point for the WMF is that the case as it stands is between two parties each of whom claim copyright in the picture: PETA on behalf of the monkey, and the photographer who claims the copyright under UK law. Whoever wins, the likelihood is that the result will be a legal precedent acknowledging that one party or the other holds a copyright. That presumably is what the WMF could avoid by filing an amicus brief, but I Am Not A Lawyer and the WMF employs several, hence the question to Lila. Rogol Domedonfors (talk) 21:28, 23 September 2015 (UTC)

Disclaimer: I'm not a lawyer either, though I've read some copyright law and court cases. I don't think it's true that "Whoever wins, the likelihood is that the result will be a legal precedent acknowledging that one party or the other holds a copyright.". The court could conceivably decide that PETA has no "standing" to sue on behalf of the macaque, and end the case right there. Or the court could decide that there's not enough evidence that an animal can reach the legal standard for artistic creation (I tend to think that even if apes have some self-awareness, they don't understand the concept of taking a photograph). Or decide that as a simple matter of law, only humans can hold copyright, without reaching the question regarding the specific facts around these images. I don't see much of a WMF interest which would justify spending the money and resources for a brief. Even if the court decided that a non-human could have a copyright, that would just mean that Commons would need to remove the images as infringements - I don't see any argument for significant damages. -- Seth Finkelstein (talk) 23:19, 23 September 2015 (UTC)

All of which supports the contention that the WMF lawyers should examine the case and decide what they need to do to make the outcome one which supports Commons. Hence my question to Lila. Rogol Domedonfors (talk) 06:24, 24 September 2015 (UTC)

Just as a remark for discussion on the topic, I strongly doubt the WMF will do anything at this time. It's too early. We don't know if the case is going to proceed, or will fail at a preliminary stage on various legal-technical grounds. And objectively, there isn't that much at stake for Commons here. Rhetorical question - How many "monkey selfies" does it have? -- Seth Finkelstein (talk) 04:58, 25 September 2015 (UTC)

Cookies in the EU

I just would like to ask when the WMF will comply with the European directive to ask users for express consent to accept cookies in its projects? – Thanks.--Aschmidt (talk) 22:49, 25 September 2015 (UTC)

Thank you Aschmidt. We are aware of the issue and -- in fact - have been looking at privacy, identity and usage data implications globally and the right solution. We hope to have a solution this year, but I think we owe the community a public discussion/feedback before we implement it. LilaTretikov (WMF) (talk) 04:36, 26 September 2015 (UTC)

Flow update -- action requested

Lila, a month ago you stated [4] that you has asked the Director of Editing to respond to a question on the current status of Flow. A fortnight later I reminded him of that request here at Meta and at Mediawiki. Nothing has been forthcoming since your original statement here. I know that he has had the opportunity to see that reminder [5], so his failure to comply with, or even acknowledge, your request is presumably deliberate: it is lacking in respect to the community and to you. Please would you take appropriate the appropriate steps to ensure that the community is given the response you wanted. Rogol Domedonfors (talk) 09:15, 23 August 2015 (UTC)

There was an update at Wikimania. And another update coming at the end of Q1. While we are not able to respond to all individual inquires ad-hoc (if we did we wouldn't be able to code anything) we have scheduled updates and goal setting that happens on a quarterly basis. LilaTretikov (WMF) (talk) 17:33, 24 August 2015 (UTC)

For the large majority of us who have not the time nor the money to attend Wikimania, can you please post that update in an appropriate on-wiki venue? If there was a significant update given it should have been made available to all. Furthermore can you confirm whether you agree with Doc James that Flow will need to get community consensus before it is implemented? BethNaught (talk) 20:10, 24 August 2015 (UTC)

• Lila, your answer is beside the point. The diff I gave points to a specific commitment that you made after Wikimania to have an update here from your Director of Editing, which has not happened after a month. My request is not an ad hoc enquiry but a request to you to honour that commitment. If you have decided not to do so, please just say so. The current quarterly goals for Flow have not been published. Responding to legitimate requests for information that should have been, but is not, made available to the community is not preventing coding work, it is work that could and should be done by the staff who are being paid specifically to engage with the community. Rogol Domedonfors (talk) 20:37, 24 August 2015 (UTC)
  • Regarding goals, almost every team has quarterly goals published on mw:Wikimedia Engineering/2015-16 Q1 Goals. The goals for the Collaboration Team are there, and have been for several months. If you have comments, I would suggest reaching out to the Collaboration Team's product manager, DannyH, who would be happy to answer any questions you have. --Dan Garry, Wikimedia Foundation (talk) 05:58, 26 August 2015 (UTC)

Thank you for that correction. I should have stated the roadmap for Flow does not have current quarterly goals I not think it affects my point. I do not propose to wander off to further venues asking yet further questions of further people when Lila has already designated a named person to respond here on this page to resolve a discussion with her that has been going on here since March. I am asking Lila to ensure that that happens. Can you help? Rogol Domedonfors (talk) 07:10, 26 August 2015 (UTC)

Update: that roadmap was brought up to date today [6], after that comment. Rogol Domedonfors (talk) 21:39, 26 August 2015 (UTC) Further update: but still not at w:en:Wikipedia:Flow. It would be nice to have a consistent, up-to-date portal for all these discussions. Rogol Domedonfors (talk) 08:29, 31 August 2015 (UTC)

Lila, just to say that in the light of recent pronouncements, already being discussed below, I regard the specific issue I raised here as closed. You may wish to pursue internally the question of why a member of staff failed to deliver on a commitment you made to the community, but I suggest that that is now a purely internal matter. Rogol Domedonfors (talk) 21:02, 6 September 2015 (UTC)

• This is interesting. It seems the WMF has relented with regard to Flow and has finally decided to take a more collaborative approach with the community. Of course, what form this work will take will be important and there is no indication how it will work out. However, the fact that, even if fully a year after the Media Viewer debacle, the WMF has decided not to force a massive breaking software change on the communities, is something to be welcomed. BethNaught (talk) 22:34, 1 September 2015 (UTC)

Spending greater energy working to improve what we have now in small incremental steps I agree is a simpler approach and one that is more likely to bear fruit.

There is a lot we can do to improve our current talk system such as 1) allowing the watching of specific sections and 2) allowing notifications between different languages now that we have SUL. Both these have a great deal of support. Doc James (talk · contribs · email) 01:28, 2 September 2015 (UTC)

BethNaught, sadly, you should know by now not to make assumptions like that.

Pinging Doc James to note this update: I checked with the Flow project manager. There has been no change - they are still continuing full speed ahead on Flow development. They are merely focused on a particular set of Flow development features.

They are working on "workflow" features, which is in theory a great, much wanted tool. It would let us build automated processes for stuff the community currently does with scripts and bots. For example to automate the AFD process we would define: Step 1 automatically put the AFD tag on the article, step 2 automatically create a pre-formatted AFD discussion page, step 3 add the discussion page to the daily list of all AFD discussions. A lovely tool making it super-easy to create that for AFD, and more. In addition, it would add proper support for making the automated-AFD-process available to everyone, without installing custom scripts. But the WMF won't built it - not for our existing pages. In my opinion.... it's rather obnoxious for the WMF to "answer" the Flow controversy by making a much-wanted tool deliberately incompatible with existing pages.

Every board member elected in the last election had a specific platform that Flow not be deployed unless the community wants it. *IF* it turns out that the community doesn't want Flow, then all of the work on the Flow-only-workflow is a dead duck. Or.... the WMF could build the much wanted workflow system to replace kludges like the Twinkle-AFD script, a workflow system that is compatible with existing pages (and presumably compatible with Flow pages). Alsee (talk) 16:04, 6 September 2015 (UTC)

This change was made [7]. From what I understand they are going to be working to improve the current / old talk page system. Doc James (talk · contribs · email) 03:00, 8 September 2015 (UTC)

Doc James, yes I saw that edit. Then a community member went to the WMF's own Flow page and made the same edit changing "will" to "may". Well.... Apparently the Flow Project Manager did not like that edit. He deleted the "may".[8]

So the current situation is that the WMF's public position on Flow posted on EnWiki says "may", while the WMF's internal page doesn't. Half hour after the WMF changed the EnWiki page from "Flow will" to "Flow may", they edited a grossly misleading explanation of the situation onto the Talk page.[9] If you read that, it sounds great. There are people at Village Pump making proposals for new projects, with passing mention that maybe the WMF pick it up now that they're not working on Flow, now that the WMF is working on stuff for the community. Except it's not true. (new edit>) Your "understand(ing) they are going to be working to improve the current / old talk page system" is incorrect. You were misled. I went to the Flow project manager and asked.[10] The Flow team is still going full speed developing Flow functionality. As I noted above, they are working on a project to replicate community-built functionality like Twinkle, but they are deliberately designing it so it won't work on our existing pages. They are designing it so it won't work with our current editor. They are designing it so that workflows like AFD can only be preformed on Flow pages, designing it so that anyone using our current editor could not participate. The WMF could design the new project to work with existing pages and the existing editor, but they won't. New development is Flow-only. If we don't convert all our pages to Flow then we get squat.

Around a year ago I tried pasting some recent Talk page article work into Flow. Flow mangled the rendering to the point of unreadability. A bunch of us came to Lila trying to explain that Flow is a chatboard and doesn't support the work we do. Eventually she agreed, she told us they'd stop to review the issue and report back to us. Well, there was no stop. The issue was never addressed. We never got any answer. They ignored the fact that Flow is not a viable platform for article work, they just keep going. After a year of development there has been no improvement. In fact they've somehow managed to make Flow worse - a LOT worse. I tried pasting the same sample into Flow. Flow mangled it to unreadability, AND Flow had multiple parsing errors, AND the Flow editing interface itself broke. Now the WMF has officially stopped any work directed to improving Flow itself, to try and make AFD into a Flow-only workflow. Alsee (talk) 03:33, 9 September 2015 (UTC)

Hi all, so just to reiterate what is happening:

1. Thanks to design research and all of you who have contributed your thoughts we identified major gaps in Flow-for-Article-Talk-Pages. It took some time to evaluate our strategy -- a process we started last fall.
2. We believe Flow won't be able to address some of those gaps exclusively (because it is a discussion board and not a curation/editing tool) and we need other specialized tools to fill those gaps, therefore we decided to focus on those gaps and build better tools for them (we don't know if we will integrate them into the current TP or into other tools such as Echo or Revision pages or create a specialized tool -- all of those possibilities are on the table).
3. We will need your help making sure we capture all of those use cases. Collaboration page is listing some here, please make sure to flag yours.
4. We do have use for Flow because many places on-wikis need a simple discussion board (like this page for example) that can benefit from what it already delivers, we have many communities asking for it.
5. For those use cases we plan to keep fixing bugs and test for usability. I, for one, plan to dogfood Flow so I can experience it first hand, like I did with VE.

Net-net -- did we listen to feedback: yes. Did we modify our path: yes. Are we throwing away Flow as a discussion board: no. LilaTretikov (WMF) (talk) 21:45, 10 September 2015 (UTC)

Lila,

• Your central point about specialized tools comes across sounding naive, in a corporate-thinking sort of way; it puts me in mind of Dilbert. "We discovered this specialized tool wouldn't do nearly as much as a general tool does; who'd have thought? So we're going to fix the problem by building lots of specialized tools." The stuff about use cases sounds like self-justifying methodology.
• If you neglect improvements to a system for years, and allow even the parts of it that work to degrade as a result of work on your pet projects, then eventually the system you want to "upgrade" will become so hard to use that people will "welcome" your replacement, no matter how badly designed it is. That's not a good reflection on the new thing you've created, it's just a bad reflection on your policy.
• You seem to imply you mean to convert this page from a talk page to a Flow page. Sounds like a way to stifle criticism of Flow here.

--Pi zero (talk) 20:44, 11 September 2015 (UTC)

I didn't say *only* specialize tools. I said *some* specialized* tools. I think everything-is-a-unique-showflake is just as bad one-size-fits-all, good products provide balance. This is why we focus on audiences and try to understand how people use software and how they work together. I don't follow your "pet project" line. On the Flow, I am not implying, I am saying I will test it in production -- not to stifle -- but to encourage testing and productive criticism. That is what "dogfooding" means -- I want to feel the pain or joy first hand ;) LilaTretikov (WMF) (talk) 23:26, 11 September 2015 (UTC)

• As I said, the point about specialized tools comes across badly. What's behind it, those of us looking from the outside in can only wait and see; I get the impression you don't like the way I described it as coming across, and that much is good since it suggests our perspectives have at least something in common.
• I worry, though, when you speak in terms of a "product". Thinking in terms of products may be a sign of trying to fit the situation into a commercial model, which is wrong for wikimedia. Commercial products seek to maximize central control, including forcing users to depend as much as possible on support from the commercial enterprise, whereas the Foundation's goal should be to make itself as unimportant as possible. Building more and more complicated software that then has to be maintained is a counterproductive approach.
• By pet projects, I was refering to Flow, VisualEditor, Media Viewer, etc.
• You haven't unabiguously confirmed or denied whether you meant to suggest deploying Flow on this page; the original remark sure made it sound that way. The metaphor behind the term dogfooding carries a worrisome overtone of voluntarily eating what others will be forced to eat; and, it doesn't seem that you, personally, would be able to learn key things about the software that way because the range of things you do on-wiki is too narrow (and presumably can't help being so). However, it would clearly be deeply inappropriate to deploy Flow in a venue where criticism of Flow should be welcome — which one hopes includes this page. If this page were converted to Flow I would be inclined to walk away from it, for multiple reasons one of which is that I would understand by the move that my absence was desired.

--Pi zero (talk) 04:52, 12 September 2015 (UTC)

As Pi zero said, the WMF needs to get out of the mindset it currently seems to have with respect to Flow, VE, MV and so on, that it makes products that it then markets to the community. It should be building software the community wants. Currently on English Wikipedia the WMF is proposing to activate VE for all new accounts and opt-in inactive editors as well, because not opting them in would cause database strain. This looks like a deplorable attempt to force VE on existing editors by the back door. Instead of a VE pet project, how about fixing the MediaWiki preferences system so letting people maintain their settings without WMF interference does not cause sever overload? BethNaught (talk) 10:35, 12 September 2015 (UTC)

Hi Beth. First, there is a misunderstanding -- the team is not proposing things but is rather asking for feedback/best way to expose VE to current editors. It is a request for advice, not marketing. We always want to avoid changing the environment that editors are used to without their explicit signal. But to hear their signals, we would need to make VE accessible in some way. We are looking to simply let people know, in-software. To your other point... the products you are referring to have all been conceived before my time. And as the organization we learned a lot about how "genesis" of products needs to change, how to get community ideas captured and productized better. That is one of the key reasons for community tech team. That said, the idea behind those products is not in itself a bad one: easier, no-markup editing and discussion board software is the standard on the web today. LilaTretikov (WMF) (talk) 23:40, 12 September 2015 (UTC)

• Semantics. Making detailed, shall we say, outlined suggestions on the Village Pump (Proposals) is clearly a proposal. Of course, the fact that any feedback whatsoever is being requested does, indeed, show some improvement in the WMF's communication. (Incidentally, for non-enwikipedians, the idea is not going down well.) Besides, there are other ways to let people know: talk page messages, or adding a mass notification feature to Echo – it doesn't require, in your words, changing the environment that editors are used to without their explicit signal, which changing the meaning of the edit button, the most fundamental on the site, would do. Lastly, the discussion about markup is one which needs to be more broadly had. We shouldn't make assumptions and follow the rest of the web like sheep, and Flow (not so much VE, since it still edits the markup underneath) threatens several key wiki ideas: that (barring exceptional circumstances) every version of a page should be visible for ever, for example, is an important method of accountability for all editors. BethNaught (talk) 06:59, 13 September 2015 (UTC)

Hi BethNaught, VE discussion is just that -- one option team is considering (I think we have old conventions that we may be stepping on, but we intend it as a route to vet some thinking/ideas). Coincidentally we just had a discussion about that and I think this current option will be scrapped even with as little feedback as there is now; the team is working on thinking through how to accommodate for real concerns and test UX against giving users understanding and choices. To your point on Echo -- this is exactly what part of the current Flow team will refocus to do. Finally, I do think we are different and we should not *blindly* follow the web, but we should also not *blindly* ignore it as we get comfortable in our version of it. Can you give me more info on threatening key wiki ideas (you mentioned history and I will check into that). What else? LilaTretikov (WMF) (talk) 18:45, 14 September 2015 (UTC)

Wnt's post here (at 21:55 4 September) speaks eloquently as regards wiki ideas. I encourage you to read it, but in summary: every past version accessible, every page has a history, edits should be reversible, archives should be browsable and not fall off into infinite scroll. These are all key ideas for editor accountability and for maintaining the historical record in an accessible manner. BethNaught (talk) 21:54, 14 September 2015 (UTC)

Lila, I see with sorrow the clear statement you have now made of your position regarding adoption of Flow on this page, here. It's one thing to recognize in theory that the seductively self-justifying WMF mindset could blind you to the inappropriateness of such an action, even after it's been explicitly pointed out to you; seeing it in practice is still saddening. I might otherwise have been inclined to wait a while to see if you respond to my most recent comment above, but given your declared plans, I post this comment here before you shut me out. Perhaps we'll communicate in future in some other venue (though given the growing atmosphere of censorship surrounding the Foundation, perhaps not). --Pi zero (talk) 22:43, 12 September 2015 (UTC)

Hi Pi, I do not want to shut you out and I really hope you will find it in your heart to try out the new layout :) My goal is to simply test and understand the current functionality and if it does a better (or worse) job than this page is currently -- for example finding and tacking comments is both time consuming and cumbersome (I missed your comment above until now and it has be hard to manage this page as some pointed out). If it proves to be worse, then I have an option to go back. I really really hope to see you try it with me. LilaTretikov (WMF) (talk) 23:40, 12 September 2015 (UTC)

LilaTretikov (WMF), there are a number of issues I want to address, but for now I can only manage this one. You can't really dogfood Flow because you don't edit. It's like someone trying to dogfood a software development environment, when they just use it to write email. You're also not the target demographic. But that aside...

Around a year ago I tested Flow. It immediately crapped out when I tried to paste article work into it. I knew it didn't work, so I never really touched it again. Well recently I have been trying it again. I tried it and, somehow, after a year of development it got a LOT worse, it exploded when I tested it. And I ran around trying to file a bug report, and your team had removed every damn Talk page for Flow, every damn Talk page for the Collaboration Team. I couldn't find a single official contact point where your team didn't FORCE me to use Flow to try to submit a bug report. And posting the bug report *in* Flow made Flow explode.

So now I've been talking to your Team about Flow, I've been FORCED to use Flow, and I did some more testing. I now have at least two bugs open at Phabricator, and god-knows how many more that will need to be opened. I wish we had one of those User Testing videos you make for research. I just had the most unbelievably horrid experience. If you could watch it, you'd be cringing. I finally tried to actually do stuff in Flow, writing a substantial message trying to submit some User Experience and Bug report feedback and review of testing I had done. And I started running into problems faster than I could write them down. I ran into glitches, the editing interface went to war with me, copy-pasting was mangling content. I've got a half written message for DannyH, but I have to take some aspirin and I'll have to do it tomorrow.

I don't like the idea of getting rid of Talk pages.... but this is completely unrelated. This is the first time I'm really using Flow, I've already got two Phabricators open and I just had such a miserable experience I don't know whether to scream or cry. Alsee (talk) 04:57, 13 September 2015 (UTC)

I'd like to correct inaccuracies: I do edit quite a bit, but I do so under IP almost exclusively; I work mostly in science. literature and medicine. The big disadvantage of IP editing is lack of VE, and specifically citoid -- so occasionally you will see edits under this handle. Also, you can communicate with the Flow team in many forums, including this one, but the most effective one is Phabricator. Can you please send links to the issues you've filed, I's like to know what is seriously broken. Thanks! LilaTretikov (WMF) (talk) 06:28, 13 September 2015 (UTC)

Lila, glad to hear you're trying it out as a contributor. I urge you to register a new pseudonymous unlinked account and try conributing as a novice to to some of the bigger projects and also try reporting or asking about technical issues with some of your staff -- the experience may surprise you. You further wrote you can communicate with the Flow team in many forums and this is part of the problem. If we had fewer portals to visit but more effective monitoring and reaction at each, then things would run more smoothly -- we should have just one on Meta for strategic coordination, planning and consultation; one on MediaWiki for technical discussion and decisions-making; and link into Phabricator for tasking; with pointers to those portals fro designated pages on the other projects.

But I may be repeating myself on those issues, and the point I really wanted to address was dog-food. Unfortunately while you eat it here on this page, you necessarily feed it to the rest of us. Now it so happens that my particular computer system does not allow me to use Flow -- for some reason, it just does not work: probably because I have a rather antiquated setup. So I shall be deprived of the pleasure of discussing these, or indeed any other, issues with you here. Unfortunate, but there it is. Rogol Domedonfors (talk) 21:16, 13 September 2015 (UTC)

Just to update you: I have actually succeeded in using Flow, once, so I need to modify my assessment slightly. It seems that I cannot usually ad a new topic to a new page, I can sometimes add a topic to a page with other topics, I usually cannot edit my own or other peoples' comments. So when you convert your talk page to Flow, I may or may not be able to carry on our discussions. I suppose it's my fault for having a set up so old that my browser support is Grade C. Rogol Domedonfors (talk) 06:29, 21 September 2015 (UTC)

I'm sorry for saying you don't edit, but if you IP edit there's not much difference. We rarely try to talk to IPs for two reasons. #1 it's often pointless, they probably won't get the message anyway. #2 if someone stays an IP then they are actively avoiding engagement. IP's have no stable identity, you can't have an extended conversation with them, they all merge into a single Persona.

The very first time I tried to test Flow, Flow completely mangled a table I tried to paste in. Column headers were gone, table cells were scrambled, and sentences from outside the table were rendered inside a table cell. Furthermore the editing interface went haywire - it became effectively impossible to do basic text editing on the content. Quiddy has partially replicated the issue, in one mode the table is mangled, in the other mode it worked for him. I don't know why, but it failed for me in both modes. I don't specifically know if there's a Phab open on it yet, but I'd assume so.

Then I tried testing something else in Flow and I know it resulted in two Phabs open. When I save a test, the show.hide template was locked in SHOW mode. At first I thought the problem was Flow simply locked show/hide as SHOW pornographic images. It turns out the problem goes away on page reload. It's on Phab now. Anyway, that test uncovered a far more serious bug...

https://phabricator.wikimedia.org/T112338 Looks like this is scheduled for the current sprint. Thank you for reporting it! LilaTretikov (WMF) (talk) 18:35, 14 September 2015 (UTC)

Someone else edited my test post. When I looked at history-view, the history didn't match my post as I originally saw it (I didn't know the new load fixed the problem). So I tested UNDO. UNDO is broken. Flow is an awful kludge were it tries to embed multiple incompatible content semantics into the same post (which raises other issues). In this particular case on reverting an edit, Flow gets confused handling the multiple semantics. Images are gone, links are gone, templates are gone. When I tried a basic revert, Flow completely destroyed my original post.

Then I tested Flow by replicating a typical Talk discussion in flow. I made notes on a number of issues. Just one of the issues is that the Flow threading model failed hard. Flow threading model fine for simple conversions, but it does not support the complex multiperson discussions we have. You can't use top posting for new replies. It doesn't work. It scrambles the posts and the result is incomprehensible spaghetti.

Every time I test I turn up catastrophic problems. I tried to write up a report on that latest testing, and I had new problems faster than I could write them down. EDIT LINK Popups doesn't want to go away. Basic copy-paste is mangling content (this may be why I can't put Tables in Flow at all). Maybe PEBKAC, but trying to use visual editing was a nightmare. The the damn "smart code" to manage invisible-format-codes for me kept doing the wrong thing. It wouldn't let me extend a line in non-italics mode, then it wouldn't let me get out of bulletpoints mode. I seriously started to hate Visual. (I guess I could have switched </> mode, but I was simply trying to use Flow as it presented in default mode.) And all this happened when I was trying to write a simple report on Flow problems. Picture one of those User Experience research videos where you record them trying out the system. Picture that user helplessly raging at the screen, while the system kept doing the wrong thing. Picture that user giving up and walking off to get some aspirin.

All of this reporting is good-will work. I feel like I'm helping you heal an injured mosquito. (Mosquitoes literally give me scars.) The project direction pre-dates you. The WMF decides to build something, goes ahead and builds it, tells the community that it's being deployed, and then asks individuals to submit bugreports and bells&whistles to hang on it. When we say the initial direction was wrong, that doesn't make us your enemy, that doesn't make us obstructionist change-averse Luddites. Opposing progress-in-a-wrong-direction does not equal opposing progress.

I know you're trying to do the right thing. I know you're trying to build towards a better future. You're taking what you've got and you're trying to make it work. You're trying to make good tradeoffs trying to satisfy multiple conflicting goals. As a programmer, as a User, the way Flow integrates incompatible semantics is an unholy kludge. Just to cite an example, what are the semantics for colon star and hash? You have one semantic where they all mean indent-1 and possibly add bulletpoint or numbered list. They can be nested in arbitrary order to arbitrary depth. (This is the textural and functional semantics of core Article pages.) And now you have an entirely different semantics where colon-space means huge indent, star-space is huge indent bullet, and hash-space is huge indent list. And what does :*# mean? Nothing, because the second set of semantics has no support for nesting them. That is not simpler, that is not easier to learn, that is a mess. The mess gets worse when we routinely copy-paste anything from anywhere to anywhere. What do we do on User Talk pages? Discuss article work. What do we do on Administration pages? Discuss abusive edits made anywhere. What do we do on deletion boards? Discuss article content. What do we do at Village Pump? Discuss proposals relating to content and functionality from anywhere on Wikipedia. We don't have "simple discussion boards". Ordinary sentences dominate, but there's an expectation that we can paste a sample of whatever we are discussing. There's an expectation that it will "just work" and "work perfectly".

My final point here is that I can't believe Flow is still viewed as matter for strictly internal review, and for merely announcing those internal plans. When all of the elected board members explicitly said Flow deployment was subject to Community acceptance, the first thing I would have done as WMF Director is take an acute interest in the external view on Flow. I don't know what the Community Consensus view is, and I don't think you do either. Alsee (talk) 23:25, 13 September 2015 (UTC)

Hi Alsee, no worries -- I edit not for glory, but to understand the experience first-hand. I do think IP edits count for quite a bit though, about 20% of WP, so can't discount those.

So sorry about the mosquito injuries, that is so painful. Thank you for helping us get the software to where it needs to be. It hurts to hear that we may have broken some things this year, let's fix that. This is why I want to really see this and experience it first hand. I definitely don't think big things like discussion boards should just be "announced" with a magic fanfare from the WMF, they need a lot of collaboration (pun intended). Optimally we would have evolved software incrementally, but we are where we are and I think we can make the best of it if we can do it *together*. As you seem to have lots of experience with it -- no software is perfect and it is a process of constant tuning and eventually getting it ready for prime-time.

I take your point that every place is a potential article discussion place. But not all discussions are made equal. This page is much more light-weight for example. So it is a good place to start and build up to more complicated use-cases. Let's see how much the new experience it breaks (or doesn't) this first and in interim maybe you and I can collaborate on some tech article and test that use case as well. You can help my editing skills :) LilaTretikov (WMF) (talk) 18:19, 14 September 2015 (UTC)

I didn't say IP contributions weren't valued. I said if someone stays an IP then they are actively avoiding engagement. That works fine so long as they are competent, don't do anything remotely controversial, don't touch remotely controversial topics, don't care if they get reverted, and don't participate in any higher level work.

An example of my work the day before I responded to an RfC. The first thing I had to do was decipher the cryptic RfC question and add an explanatory note to the RfC. The RFC author was requesting removal of warning tags from an article. I had to figure out why warning tags had been placed on the article. I checked the RFC author's contribution history (it was a big Red Flag that he had only worked on this one article). I had to DIFF his article edits to see whether he was improving the article. I had to review the overall article for Policy compliance. I had to do a larger DIFF to see how the article had changed since the warning templates were added. I made a few minor edits to improve Policy compliance. I wrote a reply that I thought the warning templates could be removed. Then I checked which two editors had added the warning templates. I pinged them both so they could weigh in. One responded within three minutes. IPs don't do that kind of work. That was an easy-peasy example of collaboration and consensus seeking. The Real Collaboration comes when you have to deal with a contentious issue and work with people who strongly disagree with you. Real Collaboration means respecting consensus even if you disagree with the outcome.

I don't know if it was an oversight or deliberate, but you didn't comment on:

My final point here is that I can't believe Flow is still viewed as matter for strictly internal review, and for merely announcing those internal plans. When all of the elected board members explicitly said Flow deployment was subject to Community acceptance, the first thing I would have done as WMF Director is take an acute interest in the external view on Flow. I don't know what the Community Consensus view is, and I don't think you do either. Alsee (talk) 14:29, 15 September 2015 (UTC)

P.S. I was just posting some of the previous issues to DannyH, and I just hit a new one. We've got simple copy-paste mangles content... a simple revert mangles content.... Flow mixing top-posting and bottom posting mangles a discussion... and now I discover that merely hitting </> twice (to glance at the alternate view) is enough to mangle what I wrote. This is getting ridiculous. Alsee (talk) 18:48, 15 September 2015 (UTC)

P.P.S. Oh joy, and now I discover that merely clicking SAVE or EDIT can mangle things. (Although this seems to be a low-grade sort of mangling.)Alsee (talk) 19:20, 15 September 2015 (UTC)

P.P.P.S. New bug. During testing with Quiddy, some private info got caught in a test post. And guess what? OVERSIGHT doesn't work on Flow. And just to top it off, at the same time I ran into another new bug. The menu item to hide the post became inaccessible. So no Oversight, no hide. My info is stuck on display. Alsee (talk) 21:04, 15 September 2015 (UTC)

Alsee can you please provide URL/phabricator to reproduce. If what you are saying is true, I am not sure how people are using it at all. LilaTretikov (WMF) (talk) 21:27, 16 September 2015 (UTC)

Have I remarked, btw, that one of the core principles by which I'm coding my wikidialog tools is that all code should be surpassingly small simple and straightforward, so that coding errors are minimized and their consequences are also minimized? One of the great weaknesses of code written by a team is that it's really hard for a team to deliberately design its output to be maintained by a single person who doesn't have a lot of time. --Pi zero (talk) 23:41, 15 September 2015 (UTC)

Somewhat cynical humor aside, there's a very important point here, I think. Projects like Flow are fundamentally based on the premise that the Foundation should be using teams of programmers to do complicated stuff; even if successful, this only maximizes the dependency of the volunteers on continuing massive investments by Foundation programmer teams to maintain the software, while minimizing volunteer control; complex software inherently reduces flexibility for the end user (note, a commercial software concern wants to maximize customer dependency and minimize customer control, making it disastrous for the Foundation to treat the volunteers like customers). What the Foundation should be trying to do is opposite on multiple dimensions: create supremely simple software of extreme generality (which is to say, true flexibility of the sort that complexity prohibits), that requires as little programmer maintenance as can possibly be arranged, designed to maximize the ability of the volunteers to do whatever they can dream up with no structural constraints and leave the complexity to the volunteers. Complexity should always be left to the volunteers; the trick is to provide simple general facilities that empower the volunteers, and complicated software is guaranteed to not do that. The failure of the Foundation to understand why wikis are successful in the first place is the single greatest short-term threat to the sisterhood atm (which is why I'm here discussing the Foundation's initiatives rather than working on various essays about things like AGF and objectivity that I put on hold a year ago; those things are important for the long run but the Foundation is likely to make the sisterhood fail sooner). --Pi zero (talk) 17:38, 16 September 2015 (UTC)

This is a better articulation of the "threatened wiki principles" I mentioned, Lila. BethNaught (talk) 18:16, 16 September 2015 (UTC)

So, first, let's assume good faith: we want the vision to succeed, to do so we need you to succeed. We may disagree on hows/whats at times, but I hope we all are in agreement on the why: we are here to help every human share in knowledge. With that out of the way... we are dealing with a lot of history and trying to make the best of that. It is a cleanup and I cannot promise that we won't make mistakes, but I do hope you see that we are trying to be more mindful of what makes us successful. Please, do give me references to what you think makes wiki work (I am always curious if there is something I am missing). Thank you for rephrasing the principals for me -- I think they are mostly still relevant even with our massively different scale. Let me tell you about our philosophy going forward. I *always* ask our engineers to make the simplest and most modular solution. This is why we are building more and more components and APIs with the idea that you can use, combine and manipulate them however your imagination allows you to. Just this month we introduced the API for WikiData query and introducing PageViews APIs. Both are something communities have asked for repeatedly. Last year we have also broken down our monolithic team structures into much smaller teams (some are just peer programmers). Flow team just broke down into four couples that are working on components as well. This is not an overnight turn-around, so -- while we strive for the KISS (keep it simple, stupid) principle -- the complexity inevitably creeps in. LilaTretikov (WMF) (talk) 21:27, 16 September 2015 (UTC)

Hi Lila. When you say "we are bulding more and more components and APIs with the idea that you can use, combine and manipulate them however your imagination allows you to", what worries me is the idea of building more and more components and APIs. It's impossible, of course, to fully capture what's really being done in capsule form like that — and I've many years' experience in just how easy it is for words like this to obscure the difference between entirely the right thing and entirely the wrong thing. There's lots of historical precedent for a pessimistic interpretation, though. Fairly consistently for some time now, design decisions have been made that are the way a conventional programmer would do things and, correspondingly, incompatible with the way wikis work so that they further degrade the purity of the wiki concept that empowers the sisterhood.

Take, for example, the lamentable decision to choose, as a means to address the computational-expression shortfall of templates, a separate, procedural language (Lua), rather than something integrated into wiki markup — I was remarking to someone recently (as I have before, in various venues) on how a separate procedural language profoundly undermines wiki-ness, and they pointed out that this solution probably came naturally to programmers who had already been through the breakup of html into separate html/css/javascript elements. I smacked myself upside the head for not seeing that myself; of course they'd think it a natural solution if they had that background. And wiki markup is from completely the other side of that story: html was cool to write by hand in the very earliest days of the web (when I used to write it by hand), and lost its appeal because it got way too hard to write by hand and then split into pieces; and wiki markup was then hugely successful because it was really easy and reintegrated the act of writing web pages into a single language. So that splitting up wiki markup is, technologically speaking, very nearly suicide, and the more of it you do, the nearer.

And no, the complexity doesn't inevitably creep in — it creeps in if you split things up into separate parts. I'll say that again: fragmenting the system is the mistake that causes that complexity creep. The interference patterns between the separate parts creates vast, unmanageable complexity both for the programmers and, even if the programmers could handle it, for the users. The only cure is to not do that. I realize you want to believe wikidata is a huge success, and some aspects of it have been successful — but the whole could have been so much better if a profoundly more wiki-ful approach had been built into it from the ground up, and meanwhile its differentness creates that harmful interference pattern, clashing painfully by inflexible mismatches with the various sister projects (which, unlike wikidata, are based on a genuinely flexible medium). And the more such stuff the Foundation introduces, the worse things will get. --Pi zero (talk) 04:51, 17 September 2015 (UTC)

continued in section #Flow update -- action requested (arbitrary section break)

Environmental Impact Assesment

There is a discussion at en:User talk:Jimbo Wales#"values and_beliefs" about the foundations environmental impact and reporting. Does the foundation have any policy on reducing its Environmental footprint? The main contributor to foundation's environmental footprint must be the servers its running, being a top 10 website means this must be considerable. There is an old press release about using a Green Data Center in Europe (wmf:Press releases/Wikimedia Selects EvoSwitch June 2009), but there is no indication if this is still running or if we are using green energy in the other centers. I suspect the other factor would be the amount of international air travel which is known to have a very high impact.

In the run up to the en:2015 United Nations Climate Change Conference in Paris it seems a timely time to examine these issues. --Salix alba (talk) 05:14, 13 October 2015 (UTC)

• I believe that this discussion is principally in the domain of the Finance and Administration team, because they are in charge of facilities. Another issue that's relevant is the environmental impact of so much air travel from staff. I don't know how feasible it is to make improvements in WMF's environmental impact, but from a cost savings perspective it makes sense to me to keep an eye on energy usage and travel costs. I think that this general subject might be the domain of Lynette Logan in the Administration department. You can find her contact info here. --Pine^✉ 15:20, 13 October 2015 (UTC)

A little change of subject

Understandably, a lot of this talk page is occupied with questions and concerns; I have some too. I'd just like to note that the Wikimedia projects, despite our many problems, will still be here tomorrow, hopefully with more highly-active editors if the recent positive trend in that metric continues. I'd also like to point out some beautiful work of our colleagues who contributed to the in Wiki Loves Monuments campaign in Pakistan this year. An organizer of WLM from that country shared some photos on the Social Media mailing list; they'll be published in this week's Signpost here. (: --Pine^✉ 23:41, 7 October 2015 (UTC)

These are truly beautiful, Pine! Thank you for sharing. LilaTretikov (WMF) (talk) 07:25, 26 October 2015 (UTC)

Parsoid

A while ago you wondered why I was running into so many problems with Flow, and how I can get new Phabricator's opened as fast as I bother to report stuff. I think I've figured where a lot of problems are coming from. A very easy mistake was made compiling the Parasoid test set. You're accidentally using a pre-sanitized test set. You're using a sample of 160k saved article edits.[11] At first impression, that does sound like a through and representative sampling of real-world editing work. Based on the latest data I saw for parsoid-testing, the Parasoid team thinks the current quality of Parasoid is:

• Zero rate of total Parasoid failure. (Obviously wrong, if I can quote the error message it did generate.)
• 1-in-600 chance of Parasoid mangling the wikitext in a way that mangles the rendered page. (Clearly wrong, I'm pretty sure I have under a hundred Flow posts and I have this happen repeatedly.)
• 1-in-4 chance of Parasoid mangling wikitext in a way that doesn't change the rendered page. (The true rate is a far higher, it's an AWFUL experience for experienced editors, and newbies will quit because they can't figure out what-the-heck is going on when Parsoid constantly re-writes stuff.)
• 3-in-4 chance of Parasoid handling a round-trip cleanly.

The reason those figures are wildly wrong is because they didn't take into account how editing happens in the real world. They only sampled final saves. Real-world editing work is done in a repeated edit-preview cycle, ending in one final clean save. Real editing work involves typos. It involves mistakes. It involves writing half an edit, and previewing the result before building the rest. It involves newbies (and even experienced editors) typing in all sorts of strange stuff when figuring something out.

If you want find out how Parasoid works in real world use you need to test it against the full sequence of preview versions leading up to the final save. In real world use the edit-preview cycle generates a whole stack of roundtrips through Parsoid. In real world use, those rountrips are littered with typos and mismatched <tags> and all sorts of malformed wiki. In real world use, the edit-preview cycle hammers Parasoid with exactly the sort of random dirty-wikitext that Parasoid chokes on. Alsee (talk) 02:06, 12 October 2015 (UTC)

Alsee thank you. I will ask for the team to look deeper at their testing methods.

• Just noting that while I haven't tested Flow myself yet, many of us who post here will be doing so after Lila's talk page is converted to Flow, so at that time we regulars will have a try at the Flow experience in this context. (: --Pine^✉ 02:45, 13 October 2015 (UTC)

Converting this page to flow would demonstrate contempt for actual input from dissenting voices. It saddens me to have to be that blunt. --Pi zero (talk) 14:05, 13 October 2015 (UTC)

• I see it differently. If the conversion to Flow and the following attempts to use it go badly, it will do so quite publicly and we will all know about it. I think that Lila's talk page is a good test case. --Pine^✉ 15:24, 13 October 2015 (UTC)

Pardon, but that's irrelevant to the issue I raised. The act of converting it in the first place is an explicit choice to limit discussion here to those willing to be forced to use flow. If Lila goes ahead with the conversion, she might still say she's open to input from dissenters, and she might believe it too, self-deception being a powerful thing, but by the conversion she would have demonstrated (I chose that word carefully) that it's not so. --Pi zero (talk) 16:26, 13 October 2015 (UTC)

Dear, Pi zero. Try to look at it from another perspective: I am doing this specifically so this discussion can be more inclusive. Most of the time I cannot find responses and correlate them. I want to see if this helps or not. If it does not, I can move it back. Let's keep an open mind and give it a chance! And, if I had contempt for you, would I be answering at all? Much more trying to convince you to try new things with me? I don't feel that way, please know that. LilaTretikov (WMF) (talk) 11:15, 20 October 2015 (UTC)

Much of what you say is answered best by what I said before, that you were replying to. It'd be redundant to say you're missing the point (and it's a hugely important point). Here's an easy question you ask, though: "if I had contempt for you, would I be answering at all?" Yes, you would. It's part of the whole self-deception thing. --Pi zero (talk) 16:54, 23 October 2015 (UTC)

I am sincerely sorry you feel that way. LilaTretikov (WMF) (talk) 07:27, 26 October 2015 (UTC)

Flow update -- action requested (arbitrary section break)

continued from section #Flow update -- action requested

┌─────────────────────────────────┘
You don't seem to have grasped the subtext. The highly flexible, extremely general software is wikitext. It does let the volunteers create whatever structure they want. Wikis work because they have powerful general software. It allows workflows to be crafted for any situation. Your attitude that we are dealing with a lot of history... It is a cleanup is very disappointing. You imply that wikitext discussions are bad. I'm sorry, but they aren't. Yes, there are some issues, but they're not insurmountable, and it's a lot better than any straight-jacketed Flow discussion system, in terms of flexibility, creativity in workflows, and power. Your consultation with experienced users prior to Flow was a joke. It asked a completely biased question in order to further the predetermined idea of the WMF. The mw:Flow/Research#Feedback during first release was almost wholly not positive in the majority. There has been no meaningful discussion with the community, that I am aware of, on the basic wiki-philosophical premise which we have outlined. Flow cannot credibly proceed, at least not on the more traditionalist Wikipedias, without such a discussion. BethNaught (talk) 22:01, 16 September 2015 (UTC)

Hi BethNaught, I was reading your comments. You have made some very eloquent points. I went to wnt's talk page on en.wp just to thank him and ask him to copy his wiki principles on meta - I hope he takes me up on my offer. I am certainly no fan of flow - though my opinions aren't as strong as others, and I agree with a lot you have said. I just want to say you are doing everyone following this a disservice by using a condescending, derisive tone in your arguments. You, more inclusively, We all, can do a better job here in communicating our dislike even contempt of flow in a better manner. If you would read my point below, projects like Flow weren't Lila's initiatives, the research seems to have been carried out during the transition phase. Is it fair to focus all the contempt and ire on her? when the researchers, developers, communication staff behind those mistakes aren't probably going to even read this. Lila is getting the unrefined version of feedback here. I see benefit of all that you are saying, but I also know we can say it better - I think. Regards. Theo10011 (talk) 22:41, 16 September 2015 (UTC)

Hi. If I may interject for a moment here, I was reading through this discussion and I think my fellow community members are overlooking some points. If I may stress them maybe it can help the tenor of this conversation. I apologize in advance for sounding self-righteous or patronizing - it is not my intention.

1. History- Flow, VE, Media Viewer, et al - were projects started or even completed before Lila took over. She had no input on them, it wasn't under her direction. I think it's unfair to keep blaming her for any particular mindset or lack thereof, that she had no input in conceiving.
2. To her credit, I see Lila engaging here an year later, and really, making an effort to follow up. You have to admit, this is a giant improvement when it came to being heard on these issues. This wasn't the level of direct correspondence we had earlier with Sue or Erik (you can read their talk pages for comparison). I think its a good sign, don't you?
3. I don't see Lila mentioning she is going to force Flow or any feature, community opposes strongly, on wiki. I think we have all learnt from the past experiences and feedback. At the same time, I only see her points pointing to an evaluation of these things. I think she is entitled to that, to learn from past mistakes - projects like Flow represented substantial resource investment - it would be irresponsible to not re-evaluate and give them a second look, even to salvage for parts or lessons or keep on ice for future reference.
4. Too many voices - there seems to be some mixed signals or miscommunication that is being taken in bad faith - staff members not responding, or responding differently. And some here are forming conjectures and conspiracies where there is no need for them. It doesn't matter what one member of a team said, or posted months ago or what appeared in a report. You have the ultimate authority on WMF initiatives answering your concerns - give Lila her due.
5. If you have questions then I think Lila is here to answer them, if you have bugs, you should report them. If you have feedback it should go on the talk page of relevant project, future references and ideas for tech developers should be saved on meta. But let's not jumble all these things here, I see some really important, eloquent ideas and suggestions, mixed with bugs and feedback and questions. Unfortunately, they will be lost once this page is archived - save them in the appropriate place if you want to make an impact.
6. It's not helpful to use polemics and over-reactions when we want to work together and make progress on common issues. Also, constantly dragging out a past Lila played no part in, is unfair to her. Let's focus the disappointment and outrage of the past, in to something constructive.

Thanks for reading. Regards. Theo10011 (talk) 22:28, 16 September 2015 (UTC)

Hi, Theo10011. If I might offer some thoughts on your thoughts. :-)

• Everyone here knows, I suspect, that these projects predate Lila. Lila hasn't yet shown signs of grokking how unwise those projects are; she does apparently appreciate that there are problems, but she's indicated she wants to fix them, which is really not possible because, regardless of superficial manifestations of the problems with each project, ultimately the problem is rooted not in details of how the project is being conducted but with the core concept of the project.
• Increased communication is a good sign; it disturbs me to even have to say that, it seems like it should go without saying. Also, though, increased communication isn't nearly enough.
• Flow etc. have enormous momentum which Lila doesn't show any real sign of trying to stop. That goes, again, to not understanding that the core concepts of these projects are bad for the sisterhood, not just details of their implementation.
• Your point about staff members not responding etc. seems to me to illustrate one of the (many) fundamental flaws of the "assume good faith" doctrine. "Faith" is pretty much an irrelevant concept, because what people intend to do and how they go about pursuing those goals completely washes out any trivial concerns of whether they "mean well". Almost everyone means well (although there are unquestionably exceptions). There's a high likelihood that all the staff involved, the people trying to interact with them, and everybody in this conversation here, "means well", whatever that means. It's entirely possible for smart people who mean well (but may be trapped in a bubble of self-justifying worldview, which is how I'd characterize the WMF mindset) to do things that cause great harm. A relevant adage is "The road to Hell is paved with good intentions."
• A second problem with "assume good faith" is that the moment someone tries to think about whether others are doing it, they run the risk of mistaking criticism for failure to AGF. I see Lila emphasizing the importance of AGF, and I see that as evidence that Lila is being distracted by asking whether others are AGFing. My advice is to stop asking whether others are assuming good faith, and just concentrate on trying to understand what they're saying.
• For my part, I've not engaged in polemics; I haven't gotten the impression others in this section were doing so either. I prefer to understate. When I said, for example, "The failure of the Foundation to understand why wikis are successful in the first place is the single greatest short-term threat to the sisterhood atm", I was not exaggerating my position; if I hadn't previously remarked on that particular point on this page (I don't specifically recall doing so), that would likely have been because it simply wasn't previously relevant to any point I was making here.

--Pi zero (talk) 00:22, 17 September 2015 (UTC)

Greetings Pi. Thanks for taking the time to respond. Allow me to make a few more observations.

• Actually, I don't see that everyone here agrees or understands that these projects predate Lila. Your "pet project" observation related to neglect of entire systems, is underpinned by the assumption that any of these were her pet projects, or the neglect they brought on the system happened under her watch - both of these observation are entirely flawed and could simply be remedied by examining the timeline of her appointment. Apart from that I see multiple examples from BethNaught and others, where the timeline doesn't justify any culpability being shouldered on one person. I suggest a re-examination of the timeline along with a consideration for a transition period.
• I agree that communication alone won't change anything. But it is a start, hopefully a positive one.
• We might be reading too much in the momentum of things, was one of my point. You and I might agree on the negative impact of projects like flow but how about we give Lila time to formulate her own opinions? Merely not opposing something without her own evaluation and being open to a trial, is not tantamount to blindly pushing for it - it's a positive sign if she wishes to try something personally before formulating an opinion - we still reserve the right to disagree with that opinion, once formed, but let her come to her own judgement first, is what I would suggest. If 4-5 individuals commenting in unison on her page were enough to deter or change her opinions - I don't think that would be a good thing either.
• I don't agree with your view on good faith. The good faith doctrine as I always viewed it was given a lack of context and intention, to expect the best in any situation, and any person, until evidence to the contrary is presented. Well meaning intentions leading to unintentional harm is a good position to make here, but I don't think it corresponds with the good faith doctrine. If the intention the staff member sets out with to fix something and in the process end up breaking other things - AGF doctrine would be tested, but if the same staff member apologizes, or agrees to fix through those mistake - would your judgement differ from mine? Assuming the mistake in question is universally agreed upon as a negative.
• I believe Lila is understanding the criticism here, but the discussion sometimes is getting derisive. And these settings being somewhat new, I can see her attempts at asking for good faith as reasonable - to not doubt the initial intentions as good.
• I wasn't referring to you or anyone particular when I mentioned polemics. And I agree, that not understanding the success and the strength of the wiki is a big oversight, but you have to understand, that this oversight shouldn't be shouldered by one person alone - WMF as it is today is not a monolithic representation of Lila's action - our board, past executives, staff, even fellow community members shaped WMF's view as it stands today.

Thanks again. Theo10011 (talk) 01:26, 17 September 2015 (UTC)

Hi, Theo10011.

• These are pet projects of the WMF; when I said "your pet projects", note I was talking on a timescale of years. Intentionality of a corporate entity (a memetic organism) is a tricky thing to talk about, but no less real for that. The WMF maintains a bubble of perception, a worldview that's very hard to think oneself out of once one gets trapped in it. The thing that makes those projects pets is that ultimately they are driven forward at the whim of the WMF. The effect of these projects is to damage the sisterhood, and if promoted successfully enough, to destroy the sisterhood in a practical sense (I'd say there's no practical difference, for example, between the sisterhood having to close its doors or simply becoming irrelevant).
• Re giving Lila time to form her own opinions: I would think providing her with input as she goes about doing that is a major reason we're discussing things on this page.
• If a staff member tries to fix something and causes more damage, relevant concerns are fixing the damage and figuring out how to arrange things are likely to work better in the future. If there are systemic attitudes that contributed to the situation, one would like to address those; it's probably irrelevant and counterproductive to ask whether or not those attitudes amount to "bad faith" or not. (Have I been guilty of asking that counterproductive question on occasion in the past? Presumably; it really is a difficult mindset to avoid when AGF is being actively promoted.) In this case, one can go yet a layer deeper than systemic attitudes contributing to the particular situation, to realize that ultimately the WMF's choice to pursue those projects at all promotes those attitudes, as well as being counterproductive in itself.
• "[...] but the discussion is getting derisive". I don't think so. I see honesty, not derision. I've warned before that AGF can be used as an excuse to dismiss things one doesn't want to hear; in that, I'm not warning against some dunderheaded attitude that anyone with common sense could easily avoid, I'm warning against something that can be blatant but can also be extremely subtle.
• WMF's view as it stands today is a coherent, self-sustaining... illusion. Each possible challenge to it has an answer within the coherent whole, such that pushing against any one part of it engages others parts to deflect the "attack". It's hard to see how to break out of the trap; there are too many parts to it all ready to come to each other's defense, and each one would take an expert of one kind or another to refute. There's a stunningly simple reason, I think, why this well-nigh-unbreakable trap has come about. Part of the WMF's function, as the WMF sees it anyhow, is to act as a cheerleader for the movement. The WMF can only promote it's understanding of the movement. And a big part of this cheerleading/promotion is to put everything in the best possible light, very much a propaganda operation (which, at least as I see it, is intensely ironic because the single greatest good the sisterhood does in the world is to retard propaganda); so the WMF has put massive collective effort into coming up with ways to defend all the various tenets of the WMF's positions. Based on very plausible motives, the WMF has been working very hard for years to close its collective mind.

--Pi zero (talk) 13:30, 18 September 2015 (UTC)

@LilaTretikov (WMF): So what are your thoughts on these fundamental wiki-philosophical issues? BethNaught (talk) 13:14, 22 September 2015 (UTC)

Lila, I too would be most interested in your thoughts. --Pi zero (talk) 11:48, 5 October 2015 (UTC)

I think our volunteers are what makes us. Period. Without you there is no Wiki.

There are many wiki principles that have been built out over the years. I agree with these with small exception that I believe Wikipedia has transcended an encyclopedia and has become something more than an encyclopedia could ever hope for: a collection of up-to-date, constantly updated, international knowledge.

Then there are more details. Those things have to evolve over time. If we don't evolve, we die. I believe most are still as important today as they were 15 years ago and some that we interpret in ways that are holding us back. The world have moved on with evolved definition of speed, interactivity, collaboration, etc. In other words: I think our principals are still true, but we don't interpret them in the context of today. We live by what we understood them to be 15 years ago.

To address some of the baseline questions: small code -- yes, extendibility (so everyone can innovate) -- yes, reversibility -- to a point (there is baggage we have created here, not everything is possible). Leaving complexity to the community means we have a lot of complexity today. Let's simplify what we've over-complicated.

Finally on the software side: I believe in making micro changes. But that is not how we run in the past and we have some major, boil-the-ocean projects that we need to bring to conclusion. If we had a choice we would have done them more incrementally, but we cannot change the past and there is good work there that we should leverage. I think we should give our editors choice: if they want change they should choose it, if not they can stay back. For our readers, we should test and deploy what improves their experience. They tend to demand a much faster pace.

Ok, that's the short of it. Go rip it up, ask me what specifically you want to know :) LilaTretikov (WMF) (talk) 00:45, 6 October 2015 (UTC)

> we have some major, boil-the-ocean projects that we need to bring to conclusion

That'll be the sunk cost fallacy. Flow is, and always will be, unacceptable because it was designed from the outset without any regard whatsoever for community input or fundamental policy. The disregard of policy continues as of last month -- Wikipedia pages are, by definition, not (discussion) boards and they should not be referred to as such. MER-C (talk) 17:18, 6 October 2015 (UTC)

Lila, while we're bringing up words that sound innocent and become problematic when examined closely — such as the comment MER-C flags out above — here are a few others I've noticed.

• "If we don't evolve, we die." Catchy slogan, but provides a very convenient excuse to do whatever the Foundation, in its (sorry, being honest) notable lack of wisdom, wants to. Such advice applies only to those who are qualified to make intelligent decisions about how the sisterhood can and should evolve, which excludes the Foundation. I don't mean to be snide; I'm reporting a very serious problem, the magnitude of which you still appear to have not yet realized.
• "Leaving complexity to the community means we have a lot of complexity today." I don't think we're talking about the same things. The complexity I see, that I approve of, is the content of the wikis. The complexity I see that I disapprove of is the software platform. When the community produces complex software (in the negative sense), generally it's been forced to resort to that complexity by lack of elegant wiki-markup-based means for what's needed. That's the problem I've set out to address (link), with a painful awareness that if I want my efforts to succeed I have to keep it far, far away from the Foundation and its processes.
• (going back further in the section) "I think everything-is-a-unique-snowflake is just as bad one-size-fits-all". This is another of those things that sounds good until considered more closely. In relative terms — compared to what the Foundation is doing now — the Foundation should be doing almost nothing to extend the software. And the difference from nothing should be extremely carefully thought out in pretty much of an everything-is-a-unique-snowflake sort of way. There's nothing else quite like creating a simple language environment for a general system: each element of the environment is... unique.

--Pi zero (talk) 23:08, 6 October 2015 (UTC)

We can definitely take any snippet of text and interpret it in many ways, not sure that's the best use of our time. We do have projects that are indeed sunk costs, others are not, some are partially so. I mean "we" as a combined we: the multi-dimensional community that we are. And the combined we makes mistakes at times. I make mistakes, I hope you too (else you would be a bot :) Evolve, means change. It doesn't mean good or bad change: let's make it meaningful to knowledge and people. Do you think of proliferation of templates as complexity? I think we will have to disagree on developing and changing software. Someone has to do it at the end and while lots of great ideas that we can build upon come from our volunteers, those were not a new mobile app or a modern editing tool. LilaTretikov (WMF) (talk) 14:00, 7 October 2015 (UTC)

Terminology is important. When someone talks about their Wikipedia profile (or worse, wiki profile), it is safe to infer they have zero idea of what Wikipedia is or how it works. WMF employees are expected to be familiar with these things -- if not, they should be given the boot. MER-C (talk) 17:10, 7 October 2015 (UTC)

Sigh, MER-C. I sure hope you don't welcome new editors this way ;). LilaTretikov (WMF) (talk) 18:35, 7 October 2015 (UTC)

Lila, some thoughts. I expect some of this will sound harsh; but that's because the reality of the situation is harsh. I do you the credit of treating you as an intelligent person who, therefore, should be spoken to honestly (which is an unwise way to speak to someone with power, but excessive honesty has been a fault of mine as long as I can remember).

• When you say 'the multi-dimensional community that we are', I think you're missing something fundamental about the whole big picture of the situation. For everyone to make more constructive and successful decisions, improved communications are needed, yes, and that's why I'm writing on this page, but it's also vital to understand what is not possible. As a practical matter, it is realistically not possible for the Foundation to be part of the sisterhood's community. It's easy to say things like "We need to get away from thinking of the Foundation and the community as different sides", but the fact is, they are different sides, and can't not be. The structure and culture of the Foundation force this separation, at levels deeper than I've seen you recognize.

  Working as a volunteer on the sisters lately (I do so mostly on en.wn and en.wb) feels like living under siege by the Foundation. Historically, the non-wp sisters have expected to receive no positive attention from the Foundation and to be damaged from time to time by stuff the Foundation does without caring about its effects on us; if the Foundation noticed us at all it tended to have the same attitude as the most insular of Wikipedians, the ones who haven't a clue how the non-wp sisters actually work (excepting maybe Commons, which they see as an adjunct to Wikipedia) and, leveraging this ignorance, imagine that Wikipedia doesn't need its sisters. Our biggest protection, or so we perceived, was that there were bounds of unreason beyond which the Foundation wouldn't go because the Wikipedian community would object; and now we're having to admit that the Foundation doesn't care what the Wikipedian community says either, while the Wikipedian community is finding out what it's like to be dissed by the Foundation the way the other sisters have been for many years. And on the other sisters, we've gone from wishing the Foundation wouldn't ignore us, to hoping fervently the Foundation will continue to ignore us, because while the Foundation ignore us they only damage us accidentally, whereas the Foundation, in paying attention to Wikipedia, damages Wikipedia systematically (which is the number one reason I've hesitated to post here at all; better to lie low and hope we don't get noticed, except that, one, I do also care about Wikipedia, and two, if the Foundation manages to destroy Wikipedia the rest of us will be sunk too). The problem is vastly, staggeringly worse than you seem to grasp, both in how far apart the two sides are, and how unbridgeable the gap is. Just about every particular thing the Foundation could change about its behavior, at this point, to improve the outcome consists of stopping things it's doing. Including a bunch of things that you've evidently got yourself completely convinced are good for the sisterhood.

• "Do you think of proliferation of templates as complexity?" Here again, I'm not at all sure if we'd be talking about the same thing. Some templates are good, some are bad, which is also true of mainspace pages. A large number of templates is no more intrisnically bad than a large number of mainspace pages. The specifics of the templates can get warped by failure to provide a sufficiently effective markup language — and that too is true of all aspects of the wiki, not just the templates. Most of what doesn't work well about the wikis comes down to failing to strengthen wiki markup.
• "while lots of great ideas that we can build upon come from our volunteers, those were not a new mobile app or a modern editing tool." That's a good illustration of attitude that guarantees a destructive relationship between the Foundation and the community. Your comment exhibits, flagrantly, an assumption that the Foundation knows better than the community; that the Foundation should be doing things instead of the community doing them; and you choose a couple of things you approve of, define them to be right, and then circularly reason that the community fell short by not doing those things. With an attitude like that, when you use the prounoun "we" to refer collectively to the Foundation and the community, you're using "we" in much the same way that a kindergarten teacher might use "we" when telling their class what the next class activity will be.

--Pi zero (talk) 01:37, 8 October 2015 (UTC)

I think we are saying the same thing about templates: they are something good sometimes bad. But on the whole they are complex :) LilaTretikov (WMF) (talk) 11:05, 20 October 2015 (UTC)

Lila, I see a couple of pitfalls in characterizing templates as "on the whole [...] complex".

• This could be an instance of "kill the messenger". The things one wants to do are sometimes complex. To prevent this complexity from hopelessly messing up the page on which we're doing it, we introduce a simple way of invoking a more complicated rule that's kept elsewhere. No matter where the more complicated rule is kept, and in what form, it is in fact more complicated; hence, complaining that it is more complicated is, as I said, blaming the messenger for the message. Quite separate from that is the question of what form to keep the rule in. The unique best possible choice of form for the rule is wiki markup, because that's the only form that integrates the rule into the fundamental learn-by-osmosis dynamic of wiki markup. Putting the more-complicated rule in some form other than wiki markup is a really profoundly bad idea, directly ignoring the fundamental social dynamics that enable wikis to work.
• One of the most basic fallacies used to support the Foundation's attitude toward improving the platform is the claim that some non-wiki-markup thing, that the Foundation has talked itself into wanting, is "better". There are two reasons it's fallacious. On one hand, wiki markup — the excellence of which is the elephant in the room that the Foundation tries to pretend out of existence — has been neglected by the Foundation for years, so that its shortcomings have become more and more problematic; it's like deliberately not keeping a house in good repair, and then claiming it has to be torn down because it's in such bad shape. On the other hand, the Foundation having chosen a (frankly perverse) direction to push things then pretends that's the only alternative to the status quo — the status quo that the Foundation itself has been neglecting (which, in context, I'd have to call malign neglect).

--Pi zero (talk) 17:49, 23 October 2015 (UTC)

There's nothing worse than thinking of something else you wish you'd said. I'm acutely aware much of the above sounds pessimistic. The question arises, is it possible for an individual person, already immersed in the Foundation culture, to learn to grok the volunteer community? Perhaps it is, if they're able to let go of all their preconceptions (a tall order). For perspective, my own wiki experience: I'd heard of Wikipedia for several years before I registered an account; the big question in my mind was, how can this possibly work? So when I did register, I approached it with a blank slate, disregarding what I'd thought I'd learned about such projects, to slowly immerse myself in the community and pick things up by osmosis. About two years after I started on Wikipedia, I started on Wikibooks — distinctly different, and I approached it disregarding what I'd learned from Wikipedia. A year after that I started on Wikinews, drastically different again, and approached it disregarding what I'd learned from either Wikipedia or Wikibooks. Since then I've continued, year by year, to get deeper insights into all these projects and the ways they're similar and different. Not easy to do; not quick; but possible. --Pi zero (talk) 14:13, 8 October 2015 (UTC)

I'd like to talk to you about Wikibooks. I am trying to understand some of the smaller projects better. I don't think we ever chatted live, but if you are open to that, it would be great. LilaTretikov (WMF) (talk) 11:05, 20 October 2015 (UTC)

After I just said I felt uncomfortable mentioning the smaller projects to you because the Foundation systematically damages Wikipedia by paying attention to it, and the smaller projects can do without that kind of systematic damage?

By the way, it's "Wikibooks", not "WikiBooks". The sister projects don't use camel case, just as Wikipedia doesn't (it's "Wikipedia", not "WikiPedia"). These projects aren't some thing that happens to be a wiki, their wiki-ness is a smoothly integrated aspect of what they are. --Pi zero (talk) 17:56, 23 October 2015 (UTC)

Hi Pi, I asked for your help. And I am guessing you just said "no"? 07:25, 26 October 2015 (UTC)

Lila, honestly that wasn't a "no". Nor was it a "yes". It was me trying (unsuccessful, it seems) to point out that, by the way you asked, you came across as having either not read what I'd written, or not considered them even worth bothering to acknowledge. That attitude being part of what I've been trying to warn you about. --Pi zero (talk) 16:47, 1 November 2015 (UTC)

Because I both read and wanted to better acknowledge your thoughts, I have asked for a more direct conversation. Sometimes -- of course -- I may misunderstand. English is not my native tongue and my command of it is far from perfect. And that is when I ask for a more direct interchange. LilaTretikov (WMF) (talk) 19:28, 1 November 2015 (UTC)

I appreciate the difficulties that can arise with non-native language. (I've got my experiences as an English Wikinews reivewer and as a conlanger to draw on. :-) A sufficiently high-bandwidth communication can sometimes get past language-barrier problems that simply can't be overcome in any other way. It might not work as well as you're hoping, because real-time conversation isn't my strong suit — I'm at my best (such as it is) when I take my time to write something.

Perhaps you can appreciate that I'm wary, in this case, because I did just say, shortly before your invitation, that the non-Wikipedian sisters are currently wary of drawing the Foundation's attention, and in your invitation the subject you particularly mentioned was the non-Wikipedian sisters. If your interest is broader than just the non-Wikipedian sisters, then that would be more reassuring from my point of view. --Pi zero (talk) 17:00, 16 November 2015 (UTC)

In order that I not misinterpret what you say, please expound on "I think we should give our editors choice: if they want change they should choose it, if not they can stay back." Talk pages are about editing. Will you confirm that no wiki which does not want Flow will be forced to take it? VE is for editing. Will you confirm that editors will always be able to choose to disable it and that everyone will always have access to the wikitext editor. (I ask this because of this communiqué, which says how the VE preference is now "Temporarily disable the visual editor while it is in beta".) Or by "stay back" do you do mean that editors and wikis who disagree with these changes should leave Wikimedia? BethNaught (talk) 18:28, 7 October 2015 (UTC)

I meant to say we are trying to do all we can to give each editor ability to work in the way they like (example: if you want to keep editing in WT on desktop, we are setting it up so you can do that even if majority of editors were using, say, SMS on mobile to edit). When it is not possible because of collaboration (example: talk page around an article) we will try to give the smallest community possible a choice (perhaps it is all editors on that article, perhaps it is another method -- welcome your thoughts on this). The issue with large groups is that there is no way to satisfy every person, but the closer we can get to a person's comfort zone the happier they tend to be. LilaTretikov (WMF) (talk) 11:05, 20 October 2015 (UTC)

Go rip it up, ask me what specifically you want to know

Long before you joined the WMF, the WMF embarked upon a series of massive and expensive projects based upon a theory. That theory was that wikitext is a serious obstacle in the on-ramp for new editors. The theory was that getting that obstacle out of the way would result in a surge of new editors. The WMF quite rightly places a high priority on the goal of bringing in new people.

Lila, I know you strongly believe in data driven decision making. I assume you agree that when data contradicts a theory, then it is necessary to stop and carefully reconsider that theory. So here's my question: What do you think about the results of the 2015 VisualEditor's_effect_on_newly_registered_editors data? (And the 3-month results.) The Visual Editor page says the Rationale for the project is increasing the number of Wikimedia contributors. Every Wikimedia_Engineering Quarterly Goals for Visual editor opens with a recitation of Involve more people than ever before in the collaborative effort of building free knowledge. However according to the research data, Visual Editor spectacularly failed to produce any positive results. According to the research data the original theory was wrong.

My second question: Are you willing to seriously consider the possibility that the wikitext-on-ramp-obstacle theory will also turn out to be wrong in regards to Flow? Are you willing to seriously consider the possibility that Flow also has exactly zero value for on-ramping new editors? Alsee (talk) 11:04, 9 October 2015 (UTC)

Lila? It's been a week over two weeks. You invited questions. Any response? Alsee (talk) 01:07, 16 October 2015 (UTC) Alsee (talk) 13:10, 24 October 2015 (UTC)

Hi Alsee, discussions boards are not so much about onramp as allowing people to connect and communicate. A huge percentage of our editors communicate off-wiki on other platforms. Why? Every time I ask they say they would rather stay on-wiki, but they do not. Why? Those are the questions we are tackling. LilaTretikov (WMF) (talk) 07:22, 26 October 2015 (UTC)

Maybe because the en:Wikipedia:NOTAFORUM rules don't allow the types of conversations that they want to have? When they try to initiate unwelcome discussions on-wiki they are blocked from editing, sometimes even on their own talk pages, because they are en:Wikipedia:NOT HERE to build an encyclopedia. Such editors will find other venues. Not to mention those who want to harass their adversaries without facing on-wiki consequences. And I don't really understand the need for email lists for public discussions that could be held on-wiki. I don't follow email lists much. Wbm1058 (talk) 14:33, 30 October 2015 (UTC)

Hmm, if that was the case, than there would not be so many volunteers using Facebook for conversations that are welcome even according to our rules. Rules are always worth revisiting, but I think the issue is deeper and has to do with usability and ease of use. Our challenge is to balance between needs of those who look for flexibility vs. simplicity. Because we have volunteers looking for both. LilaTretikov (WMF) (talk) 14:59, 30 October 2015 (UTC)

Do we have survey data regarding why people may use FB instead of WP for WP related conversations?

IMO one of the main reasons people use FB for WP conversations is because they are using FB for many other aspects of their lives. They can chat about WP will keeping up to date about their friends for example. It is not necessarily because FB is easy to edit (I in fact find it isn't easier and do not use the site).

Either way we need to look for evidence that 1) a change is needed and 2) that the change we are proposing will improve the situation. Doc James (talk · contribs · email) 18:30, 30 October 2015 (UTC)

We always look for evidence and improvement! Otherwise we will pull a feature back after the test. In fact, we iterated a few times with VE since March because the results were not good enough. Then we moved ahead with incremental releases upgrading from 5% to 100% over about 8 weeks. All the while measuring impact. This year we have been running a lot of surveys to better understand what communities need. Surveys are not perfect but they do help us reach a wider audience. LilaTretikov (WMF) (talk) 20:37, 30 October 2015 (UTC)

• I'm of the view that people having off-wiki conversations is fine, whether on Facebook, on IRC, in person, over email, on Hangouts, or any other venue that they choose (as long as they're not canvassing or doing anything else that's clearly a problem). Also, Wikimarkup is quite powerful and flexible, especially when combined with templates. That said, I think a more user-friendly interface for having simple discussions, particularly for newbies, might be helpful if it can be implemented in a way that has a favorable opportunity cost. I'm just thinking out loud here: would it be possible to have Flow set up such that individual sections of talk pages, rather than entire pages, can be structured as Flow conversations, so that Flow and non-Flow sections can coexist on the same page? --Pine^✉ 20:30, 30 October 2015 (UTC)

Pretty much anything can be done, just may be costly. I think one possibility is having two sections or two tubs. We did not want to split user experience originally with flow, but we will do whatever impacts our communities most positively. LilaTretikov (WMF) (talk) 20:40, 30 October 2015 (UTC)

So why can VE simply not be used on talk pages? I know some have said it is problems with ":" but how much of a problem is that really? Than rather than three ways of editing WP we would stick with two. Doc James (talk · contribs · email) 19:11, 1 November 2015 (UTC)

Great question. It is already embedded into the new version of the discussion page. That is one reason I want to check it out here. LilaTretikov (WMF) (talk) 19:35, 1 November 2015 (UTC)

I think you misunderstood Doc?? (Or maybe I misunderstood Doc or I misunderstood you?) When he said "stick with two" I think he meant simply activate VE on existing Talk pages, rather than making things more complicated and fragmented with the new Flow page type. (Which has two editing modes meaning FOUR ways of editing WP.) Alsee (talk) 22:35, 1 November 2015 (UTC)

• I'm not on Facebook, so was a bit surprised to hear that. I searched and found Cool Freaks' Wikipedia Club, is that the kind of thing you mean? "Cool Freaks' Wikipedia Club, a group of about 35,000 members that has achieved cult status, with around 35 additional splinter Facebook groups, has managed to conquer a corner of the Internet and free from the usual mudslinging, Godwin, and Caucasian adolescent boys who think feminism is a satanic conspiracy. Cool Freaks' theme is the re-posting of obscurantist Wikipedia articles, drawing members who enjoy a specific type of media consumption– going down 'Wikipedia holes'– clicking from one interwiki-link to the next in search of joy-inspiring esoterica." Are they actually discussing changes to Wikipedia over on Facebook? I've never seen an edit summary that said "per consensus of discussion on Facebook". Wbm1058 (talk) 00:30, 31 October 2015 (UTC)

Lila, I'm also confused by the Flow-rationale that "so many" editors are using Facebook for wikiwork instead of working here. I don't normally use Facebook, but I tried searching for quite a while and I couldn't find anything at all resembling that. I discussed the on-ramp theory because that's the Flow-rationale you gave us in previous discussions. In User_talk:LilaTretikov_(WMF)/Archive_6#Request_for_updates you acknowledged that Talk pages serve MANY diverse use cases... Flow cannot (and probably should not) address them all, and you explained Beginner users do need simpler on-boarding process. We need to solve for this on-ramp experience. You said basically the same thing in User_talk:LilaTretikov_(WMF)/Archive_6#Where_are_we: You acknowledged Flow deployment would be at the expense of losing features some of you consider critical, and explained the rationale for Flow was that (wikipages) is yet one new thing for a newcomer to understand and learn, which in theory should hinder their on-ramp speed. We are going to test this. The data will show if my hypothesis is false. If we're dropping that theory and rationale, with VE data contradicting it, okeydokey. But could you provide some links so we can look at this problem of so many people doing wikiwork on Facebook or elsewhere? If that's the new rationale for Flow, I'd like to consider it in a fair and informed manner. Alsee (talk) 01:27, 3 November 2015 (UTC)

Polish Wikipedia apparently had a relevant discussion recently. --HHill (talk) 07:20, 3 November 2015 (UTC)