Wikipedia talk:WikiProject Mathematics

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Nuvola apps edu mathematics.png
This is a discussion page for
WikiProject Mathematics
This page is devoted to discussions of issues relating to mathematics articles on Wikipedia. Related discussion pages include:
Please add new topics at the bottom of the page and sign your posts.
Frequently Asked Questions (FAQ)
Information.svg To view an explanation to the answer, click on the [show] link to the right of the question.


List of all archives

2009: Jan · Feb · Mar · Apr · May · Jun · Jul · Aug · Sep · Oct · Nov · Dec
2010: Jan · Feb · Mar · Apr · May · Jun · Jul · Aug · Sep · Oct · Nov · Dec
2011: Jan · Feb · Mar · Apr · May · Jun · Jul · Aug · Sep · Oct · Nov · Dec
2012: Jan · Feb · Mar · Apr · May · Jun · Jul · Aug · Sep · Oct · Nov · Dec
2013: Jan · Feb · Mar · Apr · May · Jun · Jul · Aug · Sep · Oct · Nov · Dec
2014: Jan · Feb · Mar · Apr · May · Jun · Jul · Aug · Sep · Oct · Nov · Dec

A response from WMF[edit]

Assiduous readers here will recall that, following a challenge from Jimmy Wales, there was a discussion here about what we might want from WMF in the way of mathematics rendering and editing. I sent a summary of that discussion to Jimbo Wales, who responded that he had "copied this text to the Board wiki and emailed the board (and Lila) asking them to read it" and would "personally recommend that we allocate resources to this" [1]. I have just had a discussion with Rachel diCerbo, Director of Community Engagement (Product) at WMF and have to report that she has told me that "It's one of those things that is "on the radar" but for now I don't imagine that we would be able to put it on the roadmap for the foreseeable future" [2]. It's very disappointing, especially since the proposal had Jimmy Wales's backing, and I'm sorry not to be able to deliver better news. However, I suppose that at least we know where we stand now -- WMF are not going to allocate resources to improving mathematics editing and rendering software in the near future, not are they going to include it in their planning. Rachel mentions James Forrester as "the Product Manager responsible for this area" but without resources it seems hard to see what there might be to discuss with him. In short, it does not appear that mathematics has a future here. Deltahedron (talk) 16:38, 19 August 2014 (UTC)

This is misleading, to put it mildly. Mathematics coverage continues to have a strong future at Wikimedia, but the strength of that future continues to be based on what the community of interested individuals wishes to put in. It would be good to see some people engaging with and helping the other excellent, hard-working volunteers who have toiled away to give you the tools you have today, as well as those to come. Instead, so far mw:Extension talk:Math/Roadmap merely has a copy-pasted list of passively-given complaints, without any sign that people want to actively help make changes and improve things. Code is not written by magic fairies. It's written by real people. It'd be wonderful to have some positive engagement here. Jdforrester (WMF) (talk) 16:54, 19 August 2014 (UTC)
"Misleading" is a strong term, I reject it, and you should withdraw it. I have done my best to represent the discussion as I see it and have given a link to the conversation on meta for the interested. The result of this discussion is that mathematics rendering and editing has no place in the WMF roadmap, and the WMF with over 100 employees and income in the tens of millions of dollars has not resources to allocate it, nor does it propose to include mathematics editing in its planning process (nor does it currently have a mecanism to do so). The content of the encyclopaedia is not written by magic fairies either. It's written by volunteers, many of them experts, for free. For a paid employee to come here and tell people who are contributing to the project, for free, by writing the encyclopaedia, that if they wish to be able to continue to do so, the way to do that is that they have to become unpaid developers as well, is -- unhelpful. Deltahedron (talk) 17:02, 19 August 2014 (UTC)
Having said which, let me pick up a couple of points to see if there's a way forward. Assuming that these volunteer developers do come forward, where is the locus for the discussion between what the editors and readers might want or need, and what those volunteers might want or be able to work on? Is it mw:Extension talk:Math/Roadmap? James is rather scornful of the "copy-pasted list", which I might mention was the compilation of discussions at this and similar pages, and the same list that Jimmy Wales was so enthusiastic about, as noted above. What was wrong with it? Is that the right place? If not, where is? Do WMF staff want to involve themselves in the conversation there, or is it solely between one group of vounteers and another? Is that where WMF staff will alert us to impending changes in other bits of software that might affect us, or will they use some other channel? What is the way forward, according to WMF? Deltahedron (talk) 18:13, 19 August 2014 (UTC)
@Deltahedron: I believe that there are slightly fewer than 60 software engineers at WMF, out of 208 staff overall (I could be wrong – seer the list on Foundation wiki). My team of six(-ish) people are the ones working on tools for editing – everything from formulæ to VisualEditor to hieroglyphics to citations to sheet music to… The rest work on areas like performance, discussions ("Flow"), multimedia, mobile phone, tablet and native app support, languages, and stopping the site from falling over. You're welcome to lobby senior management to increase the number of people working on editing tools so that we could dedicate some paid developer time to mathematical support, but characterising it as a straight choice between supporting and not supporting your particular editing requirement isn't entirely fair. :-)
On how to engage with fellow volunteers, there's nothing 'wrong' with copy-pasting a complaint, but it's unlikely to get things to happen compared to actually talking these things through with the people that can help you. I find that an isolated discussion on one wiki – even a wiki as important as enwiki – is often not very useful in getting stuff done. Jdforrester (WMF) (talk) 18:59, 19 August 2014 (UTC)
I am sorry to hear that WMF has such a small paid staff of engineers at its disposal. We have precisely zero paid staff here at WPM. This discussion is precisely because I did lobby senior management for more support for mathematics editing, and the result was, as described above, disappointing, if not exactly a huge surprise. Of course everyone here realises that things have to be prioritised, I really do not think you need to explain that. You have described what is not very useful to get things done. What, then, in your opinion, should we do to increase the probability of getting the thing we think need to be done? Who are these people who can help us, and where do we go to talk these things through with them? Deltahedron (talk) 19:08, 19 August 2014 (UTC)
Additional. You describe this as a "complaint". That's rather disimissive. It's a summary of a constructive discussion across three languages, in response to a specific request by Jimmy Wales and subsequently described by him as "a very helpful and concise statement of the issues and concerns". The fact that you dismiss it as a complaint suggests that perhaps you did not actually read it carefully? Deltahedron (talk) 19:13, 19 August 2014 (UTC)
@Deltahedron: I'm talking about this copy-and-paste, FWIW. I don't see "complaint" as a "dismissive" term, and I'm again disappointed that you try to twist my words like this; it's just an observation that giving people a list of demands (do this, then do this, finally do this) is forum-shopping, not real engagement with people.
There is no "why" in what you pasted, no actual discussion of what was on the roadmap to which you were posting, it was clearly written for another audience (you do realise that you posted a list of demands to WMF on a page aimed at volunteer developers, right? Surely you can imagine how demoralising/marginalising that would feel to those volunteers?)
As to the best way to engage with fellow volunteers, I can't speak for them as to what works best, I can just advise on what often doesn't work. Have you asked them how they would like you to engage with them? Jdforrester (WMF) (talk) 20:09, 19 August 2014 (UTC)
I am unable to regard this as a helpful answer. It is quite astonishing that the Product Manager, VisualEditor team, with a remit "to ensure that his "team understands what the community wants and needs, is focussed on the things that matter, and is engaging with and understood by the community" is quite ignorant of how and where to effectively engage with a group of volunteers, and is unable to suggest anyone who might be able to help. Let me just quote my question and your answer together
"Who are these people who can help us, and where do we go to talk these things through with them?"
"Have you asked them how they would like you to engage with them?"
This may strike you as amusing but it is unworthy of someone in your position. Now please may we have a considered answer to the question? Deltahedron (talk) 20:22, 19 August 2014 (UTC)
@Deltahedron: For the tenth or eleventh time, the Roadmap page to which I directed you last month is probably the best venue. For which volunteers are best to talk to, I don't know for sure, but looking at this history page suggests mw:User:Physikerwelt and mw:User:Schubi87. Your continued grossly bad faith responses like this make me sad. Jdforrester (WMF) (talk) 20:28, 19 August 2014 (UTC)
ping User:Physikerwelt so he knows about this discussion. (I guess cross-wiki pings like the link to mediawikiwiki above don't work). --Jeremyb (talk) 06:35, 22 August 2014 (UTC)
pong. I'm waiting for code review here [3]. I think we are on a good way here. I'll come back once the change is merged and the new Math rendering can be tested at beta-cluster. (I can be tested at the moment but does not work). I proposed to establish a mentioning system for volunteers (i.e. extension maintainers) to Quim Gil. I'm waiting for a reply here. I think it's essential that there is exactly one fixed point of contact on the WMF side for every extension used in production. --Physikerwelt (talk) 15:42, 22 August 2014 (UTC)
Good news, and thanks for all the effort you've been putting in on this. Deltahedron (talk) 16:00, 22 August 2014 (UTC)
"Probably"? You mean you don't know where or how volunteer effort is coordinated? Deltahedron (talk) 20:35, 19 August 2014 (UTC)
@Deltahedron: I don't know for sure, no. I'm not in a position to order volunteers around, and that includes telling them where and how to congregate. I can merely observe and intuit based on what looks sensible. Do you think instead I should try to tell volunteers that they can only participate in venues I run? That doesn't seem very wiki like or respectful. Jdforrester (WMF) (talk) 21:19, 19 August 2014 (UTC)
I would expect that you would know where and how volunteer effort is coordinated and conducted, so that you may effectively advise volunteers about what is and is not a good use of their time and energy, and so that once work is done, you can most efficiently integrate it into the code base. Restricting yourself to observing seems an ineffective way of doing your job of engaging with the community. The dichotomy posed is fallacious, but I only mention this because you asked. Deltahedron (talk) 21:32, 19 August 2014 (UTC)

Deltahedron's aggressive stance on this matter notwithstanding, the state of mathematics rendering on Wikipedia is disgraceful. It has not meaningfully advanced in the years that I have been editing, despite massive improvements in the available pool of open source solutions. If Wikipedia is to continue to be used as a serious resource for people studying technical fields, improvement of the support for LaTeX needs to be made an immediate priority. The WMF's brush-off response is also unacceptable. Foundation representatives need to get their heads out of the sand and address this pronto. Sławomir Biały (talk) 20:46, 19 August 2014 (UTC)

@Sławomir Biały: Again, I disagree that everything is WMF's responsibility. Wikimedia's software needs are and have always been met by a collective of editors, developers and sysadmins, some paid, most volunteers. The WMF is not the communities' mother, and isn't and hasn't been solely responsible for much of the software you use daily; in many cases, it's almost entirely uninvolved.
Differing areas of the software have differing levels of staff vs. volunteer involvement. For instance, I believe that all of the maths-related technology you're currently disparaging as "disgraceful" was created by volunteers, not by paid developers, and almost all of it back in the early 2000s (and you're right, its age is showing, as is the case with much of the software for Wikimedia wikis). In the other direction, despite long term requests for a rich/visual editor starting in 2003, no volunteers felt ready to take on the task, so WMF has funded that from 2011 onwards (though volunteer coders have contributed significant amounts there too).
That said, I'd be keen to understand what in your view would be us "get[ting] [our] heads out of the sand". Jdforrester (WMF) (talk) 21:19, 19 August 2014 (UTC)
WMF decided to fund visual editor, a solution for a non-existent problem, but not proper LaTeX support, a problem that has been known about for ten years? Obviously the social media aspect of the encyclopedia is a bigger priority for the foundation than the encyclopedia aspect. So I guess we have our answer. Thanks. Sławomir Biały (talk) 21:29, 19 August 2014 (UTC)
Indeed, we did our best, we put up our best case, with the best support we could muster, and we were turned down. That is the answer. Mathematics development from now on depends on volunteers appearing from somewhere (the magic pixie solution, I believe someone called it) and the staff being able to spare the time to help them integrate any code they come up with. James believes this is a "strong future" and while I'm willing to look for a way forward, if someone can find it, I also think we need to consider whether the time has come. WMF projects are clearly at a major change point, the old ways are no longer viable, and the old guard may find themselves no longer part of that strong future the WMF have planned. Deltahedron (talk) 21:43, 19 August 2014 (UTC)
I think I'm being quite restrained, given that James has seen fit to accuse me of "misleading, to put it mildly", "you try to twist my words", "Your continued grossly bad faith responses" and has described this community's constructive proposals to WMF as "complaints" and "demands". However, I am still looking for a way forward, and if, as James thinks, mw:Extension:Math/Roadmap is the place to go then, then we shall just have to go there. Deltahedron (talk) 20:56, 19 August 2014 (UTC)
Indeed. I think I would be more blunt. Anyone giving you the run-around on this issue should be seriously worried about their possible future at the foundation. Sławomir Biały (talk) 21:09, 19 August 2014 (UTC)
I've read a couple of these threads, and I don't understand exactly what is happening. Is the worry that an upcoming visual editor will make math editing much less pleasant than the status quo? Or is the worry that Wikipedia's math support is stagnating, while the rest of the world progresses? In other words, are we worried about things getting worse, or just not getting better? Mgnbar (talk) 21:48, 19 August 2014 (UTC)
I am not normally a pessimist, but I expect WMF to begin actively driving out mathematics (and computer science, physics, and engineering) editors before too long. Our articles are a small fraction of Wikipedia, and unlike lists of Pokémon they require extensive development effort, effort that could go to shiny toys like WP:Flow and WP:VisualEditor. Removing mathematics support would vastly simplify their lives and still satisfy 99% of their audience, so I doubt if it will be long before it happens. Ozob (talk) 02:02, 20 August 2014 (UTC)
@Ozob: I have no intention of removing the existing mathematics support the community have worked so long to build. Worry not! Jdforrester (WMF) (talk) 02:11, 20 August 2014 (UTC)
But if a high-priority project is blocked because it would require months of work on mathematics rendering, are you really willing to fight back? Where would your engineers find time? Look at their schedules: can one of them carve out three months (longer, if you account for their other responsibilities)? I'm sure that WMF is willing to keep mathematics support as long as it doesn't conflict with other priorities, but years of experience have shown that it is clearly unwilling to do mathematics development. Furthermore, because mathematics support is complicated, few people in the community have the technical skills, time, and interest to volunteer to lead such a large project. Sure, it's conceivable; but if I were the one responsible for mathematics rendering, then, given the other demands I have on my time, deliverable code would take years. With no support in sight, the obvious conclusion is that some day, mathematics will disappear from Wikipedia.
Put more simply: WMF doesn't care. Ozob (talk) 04:19, 20 August 2014 (UTC)
We had something like this a year or so ago when it became clear that there were issues about interaction of mathematics markup, Visual Editor and Flow. I'm far from saying that WMF were deliberately trying to break mathematics editing, but it was clearly not a high priority, even after they were made aware of the issues. The mathematics component of VE was a GSOC student project: this is the best we may now reasonably expect -- not that it's removed, but that it simply ceases to function in newly released software because there is no plan to do continue it, no effort available to support it, and that it it survives, it's on the basis of ad hoc unplanned unsupported uncoordinated volunteer effort. That is the "strong future" we have to look forward to. Deltahedron (talk) 06:36, 20 August 2014 (UTC)
@Deltahedron: Just to be very clear, there was never even a remote possibility that VisualEditor or Flow would not support formulæ. There wasn't ever an "incompatibility"; such an issue existed purely in the minds of people raising the (understandable, if entirely misplaced) concern that something might not work. Repeating the unfounded claim now, however, and especially doing so whilst painting yourself into the position of heroically saving the wiki from the evil WMF who just doesn't listen unless you scream from the roof-tops, is just FUD (or I could use less charitable terms). Please don't let yourself be tricked into saying false things like this. Jdforrester (WMF) (talk) 18:08, 21 August 2014 (UTC)
My statement was "there were issues about interaction of mathematics markup, Visual Editor and Flow". Are you saying that is false? If you wish to use the word "false", quote an assertion of mine; it is a waste of your, my, and everyone else's time to make statements that no-one ever made and then proclaim them to be false. Here's an example. You said "there was never even a remote possibility that VisualEditor or Flow would not support formulæ" That is incorrect, and to show you that it is, I refer you to [4], where Brandon Harris says "I cannot promise that there will be mathematics markeup in normal discussion comments". In other words, there was a time, namely 23:42 on 17 August 2013, when a senior member of WMF staff expressed his view that mathematics markup would not work in normal Flow discussions. You also stated that there was "never an incompatibility". That is incrrect, and it was resolved by using Parsoid, but not without bugs: see Wikipedia_talk:Flow/Archive_6#Mathematics_markup. So on 19 October 2013 there was an incompatibility. Please review your comments in the light of the evidence, and then withdraw your words "unfounded" and "false". Deltahedron (talk) 18:29, 21 August 2014 (UTC)
Additional comment on the claim "there was never even a remote possibility that VisualEditor or Flow would not support formulæ". I just tried it again at mw:Talk:Sandbox and mathematics does not work under Flow, a year after the discussions I mentioned. Deltahedron (talk) 19:50, 29 August 2014 (UTC)
@Mgnbar: Deltahedron wrote a summary of the proposal here. Speaking for myself, there is an element of both concerns. It is a definite concern that adequate support for editing mathematics will be removed from VisualEditor, since inclusion seems to be rather an afterthought entirely dependent upon the valiant efforts of one or two volunteers. But at the end of the day, I don't really think that will happen; the blow-back would be Wikipedia's own doom. The other element is the solid lack of improvement. We have for years been screaming for proper mathjax support (as summarized as the first "Specific" point of Deltahedron's summary). But instead of funding this important but mundane feature, the WMF throws its resources into developing VisualEditor, Flow, and other questionable and unimportant bits of flashy gadgetry. This clearly demonstrates the perversity of WMF's engineering priorities. Sławomir Biały (talk) 11:43, 20 August 2014 (UTC)

It is important to remember that James is not in a position to be able to set the goals of the organisation. As such, he could not reassign engineers even if he wanted to. I would recommend following James's advice to contribute to the roadmap. Failing that, you could make your case directly to Erik Möller, the Vice President of Engineering and Product Development. --Dan Garry, Wikimedia Foundation (talk) 02:31, 20 August 2014 (UTC)

Dan, we know that, and did not ask or expect him to. If you follow the links in my first posting in this section, you'll see that we have already made our case, at Jimmy Wales's suggestion, and he passed it on to the Board and to Lila. The answer was, as I quoted above, "It's one of those things that is "on the radar" but for now I don't imagine that we would be able to put it on the roadmap for the foreseeable future". So your advice, though doubtless well-meant, is somewhat behind the curve. We tried every element of what you suggest, and it failed. The question we are discussing now is, how we can move forward to sustain mathematics rendering and editing in the absence of any material support or guidance from WMF. That may not be easy -- indeed, I believe it calls into question the sustainability of mathematics content in Wikipedia. Deltahedron (talk) 06:16, 20 August 2014 (UTC)
@Sławomir Biały: I'd read that summary. (Thanks anyway.) What I really needed to read were the linked Wikipedia:Visual Editor and Wikipedia:Flow articles. The former says that classic wikitext/LaTeX editing will be available. So things won't get worse than status quo? The latter is a new discussion system. Does it use Visual Editor? With or without the classic mode?
I still don't understand the basis for comments such as "it does not appear that mathematics has a future here" or "Is it time for mathematicians to leave Wikipedia?" (made by the same editor here). Mgnbar (talk) 15:20, 20 August 2014 (UTC)
Those comments were made at a time when there was definite reason to believe that mathematics markup would not be compatible with VE and Flow, or at least, that they might have to be edited in a different and less convenient way than ordinary text. (It is my private belief that the incompatibility had not been consciously recognised by VE/Flow developers until it was raised by WPM members, but that may be incorrect.) After some discussion, it became clear that that mathematics markup editing would be available in VE, because a mathematics component for VEwas written, as a student summer project (ie, not by WMF staff). Deltahedron (talk) 16:48, 20 August 2014 (UTC)
Although the current system of displaying equations in Wikipedia is marginally better than the output of an old-school latex2html script popular in the 1990s, it has not been substantially improved essentially from the very early days. This is in marked contrast to the rest of the internet, for which there are open source solutions that are a vast improvement over what we have here. Largely due to volunteer efforts of some developers, we now have a MathJax system that kind of works. There remain significant performance issues that need to be fixed. Sławomir Biały (talk) 15:49, 20 August 2014 (UTC)
(ec) Mgnbar, technically it sounds as though previous functionality will not be deliberately broken. However, what we have may be described as limping along: I have stopped using MathJax because it is horribly slow and frequently fails to render. I have tolerated the uncomfortable truce between PNG and HTML in articles in the expectation that this would be resolved in the not-too-distant future. Already, the LaTeX supported here requires a number of inventive workarounds using overlaying of symbols to create new symbols, or creating a mosaic of PNGs. Watching the WMF's responses, it is evident that I should now not expect any fixes, only a slow degradation as the remainder of WP becomes progressively more incompatible with the existing LaTeX support. The WMF's attitude comes across as "it's not my problem". It is already clear that several of the top mathematical WP contributors are taking umbrage to this attitude.
I'm with Deltahedron on this. If we can find a place where a group of people with mathematical and scientific interests can collaborate while actually feeling welcome, I'll be there like a shot. And if this perception is widespread enough, the comments you refer to might have relevance. —Quondum 16:10, 20 August 2014 (UTC)
There is the Wikipedia:Right to fork. The database for en.wikipedia articles currently stands at around 10Gb compressed, 44Gb uncompressed for the articles, according to Wikipedia:Database_download#English-language_Wikipedia. Articles in mathematics categories are currently about 1% of the total number of 4,585,263 articles, as far as I know (I don't have a reference to hand), so presumably about 1Gb compressed, 5 Gb uncompressed (in pther words, one memory stick). One possibility would be to ask Encyclopedia of Mathematics whether they could host the articles in some holding area and then transfer them into the EoM mainspace. Another would be to find someone with a server to spare... Deltahedron (talk) 17:13, 20 August 2014 (UTC)
Portal:Mathematics states 30,412 mathematics articles currently. So just under 1%. Deltahedron (talk) 20:50, 20 August 2014 (UTC)
I've just done a scan of and old database dump from March and then there were 34,809 articles with some latex mathematical formula in them. 27% had just one equation, 50% had 5 or fewer and 25% has 17 or more. The article with the most equations was Propositional calculus with 758 formula. The total number of equation is 545,870 in all mainspace articles.--Salix alba (talk): 19:02, 21 August 2014 (UTC)
The is not only the EoM as an already existing alternative but also PlanetMath and possible Citizendium, though personally I prefer Wikipedia to them so far for various reasons.--Kmhkmh (talk) 00:19, 22 August 2014 (UTC)
I just want to point out a small fact that EoM uses MediaWiki with mathjax. In other words, "in theory", it cannot have a better math rendering support than Wikipedia. To me, Deltahedron seems too pessimistic. For instance, not every math article needs to have formulae and thus latex, and at least things are not broken (yet). I do find the wanting on the part of the foundation, though. Why can't they just deliver what editors want instead of delivering what we didn't ask, the latest example being image viewer. What I want see is not the continuing display of drama (korean soap is enough for me) but just simple very reasonable technological upgrade that puts Wikipedia on par with other sites. -- Taku (talk) 18:16, 20 August 2014 (UTC)
I cannot say why WMF have decided not to "just deliver what editors want", but that they have chosen this path is undeniable. There's a very interesting personal posting by WMF Board chair Jan-Bart de Vreede here. He makes it clear that WMF projects are at a crossroads and that WMF Board has selected Lila to take the projects in a particular direction. Bullet points describing that new direction include:
  • I hear a lot from the few and the angry. There is an argument I hear a lot: “We are the community, without us the projects would be nothing. We are the ones who got us here.” That is true, to a degree. But at the same time… we don’t want to be here…. We want to be much further along the road.
  • We want to attract new editors. They don’t have to become heavy editors, they could even contribute once in a while, as long as we get lots of them. We have to make it easy enough for anyone to contribute so that people once again feel that “anyone can edit.”
  • We need to move faster than ever before. This means we need to be tolerant of things we may not like and let experimentation happen. We also need to remove things we are attached to that don’t have wide adoption.
  • We need to act as one community, not 1,000. This means we cannot enact the wishes of a few hundred, but have to build processes that support the successes of millions.
  • All of this is going to require change, change that might not be acceptable to some of you. I hope that all of you will be a part of this next step in our evolution. But I understand that if you decide to take a wiki-break, that might be the way things have to be. Even so, you have to let the Foundation do its work and allow us all to take that next step when needed. I can only hope that your break is temporary, and that you will return when the time is right.
All of this sounds to me like a fair warning to those of us who have helped to build the encyclopaedia that unless we move in the WMF's chosen direction, which involves tolerating whatever the WMF decide we should have, we can and should leave; attracting new editors means shiny new editor experiences which may or may not support mathematics; somehow the "we cannot enact the wishes of a few hundred" (while I presume that it primarily refers to the various community votes against VE and MV) suggests that the (few and angry?) mathematics editors are going to carry even less weight, if possible, than before. So in view of this strategic overview, coupled with the WMF's rejection of the proposals submitted via Jimmy Wales, and the absence of any engagement between WMF staff and volunteer developers in mathematics, it seems not unreasonable to prepare for a graceful and productive exit as an option that may be inevitable in the not-too-distant future. Deltahedron (talk) 18:38, 20 August 2014 (UTC)
This story on the WMF's disinterest in maintaining a different important feature (the ability to find the source page of an image) is perhaps instructive about their attitude towards the preferences of existing editors and their ability to continue editing Wikipedia. —David Eppstein (talk) 19:14, 20 August 2014 (UTC)
Thanks for the link. (An analogous discussion on English Wikipedia is interesting.) When you say "find the source page", you mean the "File:..." page? I can get to it from the Media Viewer with one click. So that's two clicks total. It used to be one click. So getting to the "File:" page is less convenient than it used to be. On the other hand, it doesn't really affect my "ability to continue editing Wikipedia".
Similarly, the stated goals of Visual Editor and Flow seem noble and prudent for Wikipedia's future. (Whether the software fulfills those goals is another issue.) I've edited Wikipedia with PNGed LaTeX for 9 years. I can go another 5 years, unless it's clear that VE is going to break the PNGed LaTeX before then.
By the way, it's great that editors such as Deltahedron are engaging with the hierarchy, to make sure it stays aware of math typesetting issues. Mgnbar (talk) 03:46, 21 August 2014 (UTC)
Yes, you can get to it with one click, from a tiny little icon on the viewer display that their user studies demonstrated that users were unable to discover easily. But the link also mentions a proposed change to eliminate frames on images, and the "zoom" icon within the frame, which would make it impossible to get to the File: page for image maps (because clicking on the image itself would do something else). —David Eppstein (talk) 03:57, 21 August 2014 (UTC)
I've been following this on and off and have seen a lot of heated words but I can't see at all what over. I've seen no reports of problems with formulae and formula editing, or at least none that haven't been promptly and appropriately addressed. Nor are there changes on the horizon that will break formulae editing. Flow has been mentioned but that's months if not years from wide deployment. It cannot be judged now based on an early test version.
Formula support has been significantly enhanced in recent years. The addition off MathJax to MW is a welcome addition, and addressed the main complaint about the previous system, it's poor graphics quality. The Visual Editor's formula editor with live preview improves greatly on source editing. But though I don't have stats I suspect neither has replaced the previous system: editing source to generate PNGs still works, perfectly well for many if not most editors. It's not as if there was or is some pressing problem that needs addressing. There are things it would be nice to have, such as an entirely visual formula editor, but this cannot be a high priority as it would be a lot of work, nothing's broken now, and maths is only a small part of the encyclopaedia.--JohnBlackburnewordsdeeds 19:23, 21 August 2014 (UTC)
If I'm not mistaken is PNG rendering still the default setting and still looks ugly and out of place when used within text. Although Mathjax might work great for some, it is still experimental and afaik not without issues. Insufficient math/latex support of VE and Flow is imho a big issue as well, at least when they become implemented as default options. A system for talk pages of math/sciences/engineering articles that has no proper latex support is imho a no-go, then I rather use the current talk page as it is.
More generally speaking I'm getting more and more the impression that the WMF develops a lot of software that large parts of the community don't want or don't care for while important requests from the community seem to be put on the back burner. In addition board member statements like this one(m:User_talk:LilaTretikov#Our_Future_and_the_role_of_the_Foundation) make you wonder, whether they have their priorities and the project goals straight. They seem to primarily measure Wikipedia's success in terms of traffic rankings, in comparison to commercial internet companies and fancy guis. Such a mindset loses sight of the fact that Wikipedia is primarily about free encyclopedic high quality content and the success should be measured by the degree to which this is achieved. Traffic rankings are merely a proxy to assess that (high quality presumably leads to high traffic) and a particular stylish or fancy interface is at best an afterthought here. The notion that a state of the art gui will bring us much needed new editors willing to provide man hours and quality work seems a bit of a pipe dream and given that WMF is increasingly pissing off its current editors providing the core content is actually almost a bad joke.--Kmhkmh (talk) 22:12, 21 August 2014 (UTC)
The VE maths editor has improved since I last looked. They have addressed the problem with the edit box obscuring the equation, now it always appears just below the displayed equation. The box for the equation is a bigger, it does still require some scrolling for the more complicated expressions. Whilst it does not respect my preference for MathJax while editing, always using PNG mode, it does now gracefully go back to displaying MathJax after saving. An annoyance is that if while editing an equation there is a temporary syntax error (say if your start a \begin{align} but have not yet added the closing \end{align}) the displayed equation will show a red "failed to parse" message. This can cause the page to jump about a lot flipping between a good display and the red error message.--Salix alba (talk): 22:29, 21 August 2014 (UTC)
@Salix alba: For what it's worth, one of the team in their spare time proposed a change in May to show MathJax rather than PNGs for formulæ in VisualEditor, and editing to MathJax rather than server-provided PNGs, but it needs some work. Jdforrester (WMF) (talk) 23:14, 21 August 2014 (UTC)
Just posting to note my agreement with JohnBlackburne. --JBL (talk) 01:43, 22 August 2014 (UTC)
Despite my earlier statements in this thread, I feel compelled to defend WMF a little here. You and I may be comfortable with wikitext and with how discussion on WP is done, but many people—most especially potential new editors—aren't. I've lost count of the number of times I've seen IP editors start a badly formatted thread in the middle of a talk page. They don't do it on purpose; but they don't know how to edit, and they don't know how talk pages work, and they make good faith mistakes. I am absolutely certain that there have been many others who made it to the talk page but gave up without editing because they were overwhelmed and intimidated. Talk pages are an ad hoc collection of conventions structured around tools meant for editing articles. They are—completely objectively—terrible.
That doesn't mean that I think that VE and Flow are good ideas or will be a success, and I'm not convinced that they're a good use of development resources. But I'm also willing to accept that the Foundation might be right to try. They've invested a lot more time thinking about this than I have (I prefer to think about math), and maybe I would come to the same conclusions as them if I had spent that much time thinking about these problems.
What I worry about is what WMF will do about support for less mainstream features. That includes not only mathematics (my biggest personal concern) but also features like those that Jdforrester mentioned above (e.g., music and hieroglyphics). I am not convinced that WMF is in a position to accurately evaluate the needs of the editors who work on specialist topics, and I'm still convinced that WMF doesn't care about us. Ozob (talk) 01:45, 22 August 2014 (UTC)

I should note that MathJAX in Wikipedia (besides being opt-in) uses client-side rendering which is pretty slow in practice, something that is well known: JMP EAX (talk) 00:19, 22 August 2014 (UTC)

Why are we not doing server-side rendering? -- Taku (talk) 22:02, 22 August 2014 (UTC)
Server-side rendering is, as I understand it, the plan for the future that Physikerwelt is working on. From m:Extension:Math: "Beginning from Math 1.2 // MW 1.23+ you can use a Mathoid server that uses MathJax to convert texvc input on the server side to MathML+SVG rendering. Mathoid is the rendering mode that is going to be used on future Wikipedia." -- Jitse Niesen (talk) 15:00, 23 August 2014 (UTC)
Looking at MathJax Output formats the SVG mode is only supported by IE >=9 not all andriod devices support it. Its also limited in the character is can display so some obscure character might not work. MathML mode only works as standard on firefox and with a pluging on IE. So both server-side options will have some holes in support. The other problem with server-side is that it can't always respond well to client side changes, such as changing the page width, zooming or using different locl font size. Server-side rendering will work for a lot of people but it not the perfect solution, for that we will have to wait until the other browsers all adopt MathML (which is taking a little longer that waiting for the WMF).--Salix alba (talk): 20:32, 24 August 2014 (UTC)
  • Math 2.0 is now live in betlabs. [5] Thanks to User:Jdforrester (WMF) for the support. --Physikerwelt (talk) 21:29, 22 August 2014 (UTC)
    • Page takes forever to load in Google Chrome, i.e. nothing gets really displayed. JMP EAX (talk) 16:54, 23 August 2014 (UTC)
Same for me in FF 31.0. Deltahedron (talk) 17:12, 23 August 2014 (UTC)
    • Page is loading fine for me (FF 31/win8.1). However, first test page with MathML enabled [6] produced "(Failed to parse (MathML with SVG fallback (experimental): Invalid response ("<p>Error fetching URL: couldn't connect to host...</p>") from server "http://localhost:10042":)" error messages instead of equations.TR 11:49, 25 August 2014 (UTC)
    • The beta labs server seems to be working fine now and does give me MathML. -- Jitse Niesen (talk) 13:55, 30 August 2014 (UTC)


I have used this to covert some tikz stuff I made to SVG (though noting for Wikipedia yet.) It seems it can also convert equations to SVG via htlatex at least, but that requires some htlatex hackery/config: [7]. Beware that unlike dvipng, dvisvgm needs ghostscript to actually work (at least with tikz; that's because it relies on gs as interpreter for PS/PDF DVI specials. There's a direct SVG driver in tikz that can—in theory—produce SVG directly, but it has some major bugs. Also, that driver probably won't help with equations, which tikz itself doesn't handle.) JMP EAX (talk) 22:06, 24 August 2014 (UTC)

There are some production-quality SVG TeX backends like BaKoMa (which can even produce a bit of SVG+SMIL, i.e. declaratively animated SVGs), but they are all commercial. The htlatex/dvisvgm hackery seems the only free one, besides MathJax, which is honestly even more silly on the server side (via Mathoid) as it basically runs a fully fledged web browser process to render stuff and MathJax re-implements TeX in Javascript (and not terribly well). JMP EAX (talk) 22:10, 24 August 2014 (UTC)

mini-rant about MathJax[edit]

For some backgrond info: MathJax was basically made by Design Science (company), the makers of MathType. Before you think that's awesome, MathType was the tech behind the equation editor before Office 2007. M$ ripped out that (or rather made it legacy) after they implemented their own editor which uses the new OpenType Math fonts (and standard) that M$ designed inhouse [8], which is way, way better, coming close to TeX quality. Now MathType (Design Science) was probably going to go the Y&Y way [of the dodo] after that, but they managed to reinvent themselves as a web services company. Their actual tech is quite meeeeh, but it's the only free web thingy for math basically, given that browsers have started to drop [direct] support for MathML last year. MathJax is basically a ghetto monopoly now. JMP EAX (talk) 22:34, 24 August 2014 (UTC)

MathJax does not use actual OpenType Math font features (i.e. the superset of TeX math font metrics that M$ put into OpenType Math and recently into the ISO standard thereof). MathJax uses the STIX fonts, which themselves don't yet have that. Most glyph positioning is done by hackery in MathJax. The XITS fork of STIX (and Asana Math) have actual OpenType Math support in the free fonts world. JMP EAX (talk) 22:48, 24 August 2014 (UTC)

FYI, here's the WMF-ish plan: [9]. JMP EAX (talk) 05:06, 25 August 2014 (UTC)

MathJax alignment issue.png

And just so you don't think I'm just gratuitously ranting about this, I've attached a sample problem with MathJax rendering in Google Chrome, where MathJax actually has to do some rendering/positioning (as opposed to just translating a subset of LaTeX to MathML, which is what it does in Firefox.) It's not an optical illusion, the enlarged version also has baseline alignment problems between sub-parts of the aa-1 formula. JMP EAX (talk) 07:11, 28 August 2014 (UTC)

So, what do you suggest for us? It's all very well to rant about MathJax, but without alternative suggestions that won't improve matters. From your description, it looks like it will be a lot of effort to get the dvisvgm route to work and it is not clear that it will be better. -- Jitse Niesen (talk) 08:26, 28 August 2014 (UTC)
No one has an illusion that there is a perfect solution. Given that the web doesn't support math natively (mathml was once promising but isn't getting anywhere), there will remain some degree of hackery in a foreseeable future. What we're is asking some "modest" performance and appearance improvement that puts Wikipedia "on par" with other sites. This is a feasible cheap goal. By the way, although Deltahedron likes to frame this in a math-editing issue, this is actually a larger issue: Wikipedia as a website looks old. For example, the main page is another part of Wikipedia that has remained unchanged for "ages". WMF clearly understands this and that's why they are trying to put stuff like VE and media viewer to modernize the website Wikipedia. (Ah, this reads like a rant.) -- Taku (talk) 12:27, 28 August 2014 (UTC)
But Deltahedron did hit a nail in the sense that the fundamental problem is no one, as it seems, is accountable for math rendering; that is, it would be nice if there is some math-tech guy who fixes the math implementation here whenever it breaks and also plans to a further development (not someone "oh, I don't know it's not part of my job description.) -- Taku (talk) 12:39, 28 August 2014 (UTC)
Since you guys are probably (directly or indirectly) paying AMS dues, it might make some sense to try talk to whoever is paying for MathJax at AMS to convince them that money is better spent on native MathML support in browsers. None of the MathML implementers (in any browser [engine]) got paid anything for their MathML implementation work. Instead AMS decide to sink money a polyfill (MathJax) that will always suck in the long run. But given the disaster that the STIX fonts were, I'm not terribly surprised AMS made another bad engineering decision... JMP EAX (talk) 13:05, 28 August 2014 (UTC)
You think Microsoft will implement MathML support for some cash? And then a decade later we can finally enjoy MathML in the majority of common browser versions? ;) --Unverbluemt (talk) 12:02, 31 August 2014 (UTC)
FWIW, I've asked someone I know (former prof of mine) who is family closely associated with the inner circle of ACM SIGACT and they don't have funds for something like this, nor do they see it as much within their remit. JMP EAX (talk) 18:26, 28 August 2014 (UTC)
I am not a member of the AMS, but will explore this within the UK community. Email me if interested, as this is clearly no longer a WP or even WMF issue. Deltahedron (talk) 18:33, 29 August 2014 (UTC)

Alignment of PNGs with text[edit]

The following might be of interest JMP EAX (talk) 02:28, 25 August 2014 (UTC)

MathJax and texvc don't support the same subset of LaTeX[edit]

I just realized this when I kept seeing the flashing "failed to parse [blah, blah]". Turning MathJax off makes that error permanent. Basically MathJax supports \mkern (a fundamental TeX command) but texvc chokes on it. JMP EAX (talk) 17:00, 28 August 2014 (UTC)

I believe the current plans are to run all equations through texvc first. Basically this means the subset of LaTeX wikipedia maths supports is defined as those parts supported by texvc. There are a few other incompatibilities documented at Help:Formula#Unimplemented elements and workarounds.--Salix alba (talk): 17:17, 28 August 2014 (UTC)

Maltsev or Mal'cev or Malcev varieties or conditions[edit]

Do we really not have page for these? I tried various permutations, but I still can't find them on wiki. Nor are they mentioned in Anatoly Maltsev. JMP EAX (talk) 15:58, 31 August 2014 (UTC)

There is a brief mention in Group extension#Central extension and we have Malcev algebra and Malcev-admissible algebra, but I could find nothing else. If you would like to create an article for Malcev variety, go for it! --Mark viking (talk) 16:34, 31 August 2014 (UTC)

τ=2π in the FAQ[edit]

I never heard of this. Is it even something worth including in the FAQ? JMP EAX (talk) 22:40, 3 September 2014 (UTC)

Search this page's archives for tau and you will see many discussions over the years. After doing so, perhaps you will become a Tauist. --Mark viking (talk) 22:52, 3 September 2014 (UTC)
I see. By the way, there's an article that has been stuck in draftland for a couple of years: Draft:Tau (Proposed mathematical constant). JMP EAX (talk) 00:50, 6 September 2014 (UTC)

Discussion notification[edit]

I have posted a request to change the URL for Zentralblatt MATH linked to from WP templates and modules from HTTP to HTTPS. Please feel free to comment there. It Is Me Here t / c 16:03, 9 September 2014 (UTC)

Edward Nelson dead?[edit]

An anonymous editor has edited Edward Nelson to claim that he is no longer with us. I am very sorry to hear this, if true — he was my favorite ultrafinitist, by a lot. I corresponded with him briefly at one point while I was in grad school. Can anyone confirm or refute? --Trovatore (talk) 01:49, 11 September 2014 (UTC) PrimeHunter (talk) 02:33, 11 September 2014 (UTC)

Taylor series in several variables section[edit]

First formula - shouldn't there be a factorial of the sum of indexes instead of the product of factorials of all indexes? — Preceding unsigned comment added by (talk) 11:48, 11 September 2014 (UTC)

See Talk:Taylor series#Multi-index notation. Sławomir Biały (talk) 12:06, 11 September 2014 (UTC)
To No. When differentiating with respect to one variable, the others are held constant. So only the exponent of the one variable is brought down and needs to be cancelled out. JRSpriggs (talk) 12:29, 11 September 2014 (UTC)