Wikipedia talk:WikiProject Mathematics

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: 3

Please add new topics at the bottom of the page and sign your posts.

Is this real or a hoax?

At the risk of advertising my mathematical illiteracy, could someone take a glance at Redundant proof just be sure that it is not someone's idea of a joke? It popped up on my NPP queue and frankly if it were written in ancient Aramaic I'd have had a better shot at grasping it. Thanks... -Ad Orientem (talk) 22:23, 7 April 2014 (UTC)

I don't know anything about it so I can't be sure, but it doesn't push my "hoax" button. I suspect it's the work of a new editor adding stuff on proof compression, whatever that is exactly, and scattering it over several articles without giving sufficient context. Will need cleanup. --Trovatore (talk) 22:46, 7 April 2014 (UTC)

The sole WP:RS listed in the article has only been cited 8 times (fide google scholar). Not sure this clears the notability hurdle. Lesser Cartographies (talk) 23:08, 7 April 2014 (UTC)

Well, that could likely be addressed by consolidating the articles. I think eight cites on a highly technical topic is at least borderline, at least if they're high-quality cites, but probably better if the content isn't scattered all over the place. --Trovatore (talk) 23:41, 7 April 2014 (UTC)

Ah, good to know. In compsci systems work, my rule of thumb is >100 cites before we can hang an article on a publication. That's highly domain-specific, of course; thanks for the reminder. Lesser Cartographies (talk) 01:37, 8 April 2014 (UTC)

Redundant proof is not a joke; automated theorem provers can generate sub-proofs that are redundant with the main proof or with other sub-proofs and elimination of redundancies can be important in applications like proof-carrying code. But the editor, in addition to creating a number of articles, (Redundant proof, LowerUnits, and Resolution inference), seems to be refspamming the Fontaine, et. al. ref, or at least it seems unlikely that this one ref is the best one for several different articles. --Mark viking (talk) 00:33, 8 April 2014 (UTC)

OK it looks at least somewhat legit. As for the other details, I will leave that to the experts on this forum. Thanks for the help! -Ad Orientem (talk) 00:46, 8 April 2014 (UTC)

It looks to me more like it should be a term within some other topic rather than a topic in itself and only split out if there was more on it. Wikipedia isn't a dictionary, though we can use redirects for terms within a topic so it does actually have a dictionary aspect. Dmcq (talk) 08:24, 8 April 2014 (UTC)

As far as I understand, this article is about a specific (among several) notion of formal proofs (which one?). It uses, without any definition, a non-standard notation ($\odot$), and does not explain it relationship with the standard (in operational semantics) notation of the example (I have not found where this notation is defined in WP. On the other hand, removing redundancies in proofs is a specific instance of rewriting. All this being considered, the special case considered here is particularly minor and non-notable, and I strongly support the deletion of this article. D.Lazard (talk) 11:35, 8 April 2014 (UTC)

[Sorry about the delay]. Doesn't appear to be a hoax (although I could be wrong), but none of the articles mentioned above are standard usage. LowerUnits should be deleted entirely; redundant proof redirected to a paragraph in proof compression, if an actual use can be found; and resolution inference be redirected to resolution (logic) (with the identical section removed as being too detailed. If I have time, I'll replace the Proof compression article with something actually referenced to usable references. — Arthur Rubin (talk) 20:22, 21 April 2014 (UTC)

The notability would not be as a concept (a special case of string rewriting) but a problem or research area. It might be placed under proof compression, but that can suggest a focus on efficiency of the proof, making it as small as possible, whereas finding or eliminating the generation of redundant subproofs is a relevant concern in producing human-readable proofs. As with any other program, an automated proof's lucidity can be compromised by optimization. ᛭ LokiClock (talk) 00:17, 23 April 2014 (UTC)

TeX not rendered

Is TeX code failing to get rendered on Wikipedia today, or is it just me? Michael Hardy (talk) 18:39, 17 April 2014 (UTC)

You do not appear to be alone. Bill Cherowitzo (talk) 19:27, 17 April 2014 (UTC)

Yes, I'm seeing the source code in all its glory. RockMagnetist (talk) 19:57, 17 April 2014 (UTC)

I found them rendered, but as PNG, even though I used to have MathJax enabled. So I checked my preferences and found them set to "PNG" rather than "leave as TeX" and "MathJax" unchecked. The combination of "leave as TeX" and "MathJax" checked seems to give the status ante for me. — HHHIPPO 20:23, 17 April 2014 (UTC)

This is due to bugzilla:63915. All users who used to have mathjax enabled, need to enable it again manually atm. —TheDJ (talk • contribs) 20:59, 17 April 2014 (UTC)

@TheDJ: I am getting inconsistent results with MathJax enabled. Gamma function looks fine, but most others (e.g., Calculus) are full of [Math processing error]). RockMagnetist (talk) 21:39, 17 April 2014 (UTC)

Do you have more information ? browser, os, nageh version/core version, other installed extensions ? —TheDJ (talk • contribs) 06:25, 18 April 2014 (UTC)

@TheDJ: Mac OS 10.9.2, latest versions of Chrome, Safari and Firefox. I can't make the javascript javascript:alert(mathJax.version) work on any of them. As for extensions, do you mean Wikipedia gadgets or something else? RockMagnetist (talk) 16:37, 18 April 2014 (UTC)

I advise everyone to check: "Leave it as TeX" and the "MathJax" option. The "PNG" + "MathJax" option will take some time before it is fully working. —TheDJ (talk • contribs) 06:33, 18 April 2014 (UTC)

Isn't it a bit complex having a radio button and a checkbox in the preferences. When I reset my preferences in a rush, I didn't initially notice the subtle difference. Wouldn't it be simpler just to have a three-way radio button PNG/tex/MathJax. --Salix alba (talk): 07:21, 18 April 2014 (UTC)

The way the options are displayd is a logical fallacy; only one should be enabled. PNG + MathJax result in images loaded first before MathJax kicks in... what a waste of bandwidth! — Edokter (talk) — 12:27, 18 April 2014 (UTC)

No, it's on purpose, it will allow us to come closer to the point where mathjax is independent from how we serve the content. It's all with the long term vision of png reliable rendering, searchable svg math, mathjax/mathoid rendering for the pdf renderer etc etc. —TheDJ (talk • contribs) 14:00, 18 April 2014 (UTC)

I appreciate that the long term view is an important one, but I think there is a lesson to be learned from the fiasco of the broken align environment that the short term view cannot be ignored. If in the short run readers see something they don't expect, then there's a good chance that they will change the article to something that is actually worse when everything is finally working properly. Sławomir Biały (talk) 14:44, 18 April 2014 (UTC)

• As a matter of interest, where do we go to see the official long-term view? Deltahedron (talk) 16:11, 18 April 2014 (UTC)
  • Since support for Math is a pure volunteer driven effort, there is nothing that you could call official. And there are discussions on the wikitech-l mailing list every now and then. —TheDJ (talk • contribs) 19:02, 18 April 2014 (UTC)

I think I knew that already to be honest. Isn't it pretty disgraceful that with an income in the tens of millions, the WMF can't find any resources to allocate to support for mathematics? Deltahedron (talk) 19:11, 18 April 2014 (UTC)

That's my point. It's really not a very pleasant work to break down everything into atomic changes. With the new version of the Math extension that Gabriel Wicke and me developed in September last year MathJax is executed on the server side, which solves the performance problems. The problem with this development was that it changed from the old version to the new one in one step. I turned out that nobody from the foundation would ever find time to do a code review for such a large change. Therefore I had to break down this change into a number of small commits. This is really a lot of useless work and causes some unfortunate side effects. I'd really appreciate if the Foundation could allocate some paid developers to review the changes in the Math extension at once. But it seems that there is no money for the math extension so everything has to be developed and reviewed for free.--Physikerwelt (talk) 11:43, 19 April 2014 (UTC)

Is there a test wiki we can see this in action? I tried doing all the vagrant stuff but got lost git problems. It would be nice to see what then end results of your hard work are going to be.

I've started a thread at User talk:Jimbo Wales/Archive 161#The problem with developing maths rendering. Maybe that can help bring some foundation attention.--Salix alba (talk): 14:02, 19 April 2014 (UTC)

What are the long-term plans anyway? I use "render as PNG" since it usually works and is at least ten times faster than MathJax. Please don't tell me it's going to be MathJax only. YohanN7 (talk) 19:22, 18 April 2014 (UTC)

No. See the demo (based on the Math2.0 code from Oct. 2013) http://math2.beta.wmflabs.org/wiki/Dedekind_sum. With Firefox you see MathML and with Chrome the SVG output of clientside MathJaX rendering.

The checkbox option for the orthogonal client-side mathjax rendering option is availible for registed users in addition. --Physikerwelt (talk) 11:57, 19 April 2014 (UTC)

On this page I'm seeing nothing - lots of empty formulae - but for one error (the second of the two formulae here):

Failed to parse(MathML (with SVG fallback): Invalid response ('There was a problem during the HTTP request: 500 Internal Server Error') from server 'http://mathoid.eqiad.wmflabs:10042':): s(b,c)={\frac {1}{4c}}\sum _{{n=1}}^{{c-1}}\cot \left({\frac {\pi n}{c}}\right)\cot \left({\frac {\pi nb}{c}}\right).

This is with Safari 5.1.10 on Mac OS 10.6.8. On first reading 'That's my point...' my thought was something like this, i.e. hosting it on a separate server. That's how Lua was tested on test2. Then promote the heck out of it on here, at the Village Pump, any other math rendering related pages, and any discussion that math rendering's raised. That should given time let it be thoroughly tested and demonstrated to convince people to accept it in one go. --JohnBlackburne^words_deeds 18:27, 19 April 2014 (UTC)

And I'm seeing a mass of LaTeX source code and the same red error message (Fedora/3.6.24-1.fc14 Firefox/3.6.24). Deltahedron (talk) 18:41, 19 April 2014 (UTC)

I'm getting problems with Help:Formula not rendering with MathJax switched on. I'm getting

GET http://en.wikipedia.org/extensions/MathMenu.js 404 (Not Found)
GET http://en.wikipedia.org/extensions/MathZoom.js 404 (Not Found)  

which is odd as they are the wrong urls. I'm not sure if its some problem with my set up or a cache bug. I've narrowed down the problem a bit my sandbox does not render correctly. It would be good if someone else to verify if its a bug. --Salix alba (talk): 06:52, 21 April 2014 (UTC)

Seemed to be a problem with using MathJax as a section title. Fixed now.--Salix alba (talk): 07:51, 21 April 2014 (UTC)

See https://en.wikipedia.org/wiki/User_talk:Salix_alba#Math_demo --Physikerwelt (talk) 11:21, 21 April 2014 (UTC)

• Perhaps meta.wikimedia.org/wiki/Grants_talk:APG/Proposals/2013-2014_round2/Wikimedia_Foundation/Proposal_form might be a good place to ask WMF about their future plans. Deltahedron (talk) 18:43, 21 April 2014 (UTC)

Gabriel and me wrote a paper about the future of the Math extension. The preview is availible at arxiv. For the reference I'd like to add a link to the gerrit code review web interface and the development branch of the Math extension that already uses MathML and SVG. Physikerwelt (talk) 10:14, 25 April 2014 (UTC)

AfC submission - 17/04

This is another one that doesn't completely fall under the Mathematical domain, but it's been lingering in the queue for too long. Wikipedia talk:Articles for creation/Exponential Search. Thanks, FoCuSandLeArN (talk) 21:08, 17 April 2014 (UTC)

I'd say this one looks credible (though I didn't read it thoroughly). Is there no Wikiproject for Computer Science? They'd be able to tell (at least if it exists). YohanN7 (talk) 05:39, 21 April 2014 (UTC)

There is such a project: Wikipedia talk:WikiProject Computer science. I have it watchlisted, but it's not very active compared to this one. —David Eppstein (talk) 05:54, 21 April 2014 (UTC)

• Also Draft:Cell-probe model. FoCuSandLeArN (talk) 23:48, 20 April 2014 (UTC)

Category for discussion

Please see Category:0 hyperbolic volume knots and links and Wikipedia:Categories for discussion/Log/2014 April 17#Category:0 hyperbolic volume knots and links. The discussion is about how to name the category, not whether it should be kept as a category. —David Eppstein (talk) 21:18, 17 April 2014 (UTC)

Elementary symmetric polynomial

Hi!

I am usually a proponent of having (even detailed) proof outlines in the articles, but here, there is a proof outline, an alternative proof and a self-contained algorithmic proof of the same theorem. All of them are fat, and none of them have references. I don't want to just remove (some of) them without some form of consensus. There are probably few watchers on that page, therefore I raise the issue here. YohanN7 (talk) 21:01, 18 April 2014 (UTC)

If they are unreferenced then I would remove them as such per WP:BRD. Deltahedron (talk) 21:18, 18 April 2014 (UTC)

J-structure

J-structures are an alternative approach to developing Jordan algebras proposed by T. A. Springer in a 1973 book. There is a page J-structure which currently redirects to Jordan algebra, but the latter article does not mention J-structures at all. Curiously, there is an article hidden underneath the redirect [1] which was overwritten by Mathsci on the grounds that it had been written by a banned user [2]. The article itself seems tolerable but I prefer to get consensus before restoring anything by a banned user per Wikipedia:Banning policy. Deltahedron (talk) 21:08, 18 April 2014 (UTC)

Even banned editors can make good edits. Given the ban, Mathsci was right to revert the edit. But if you have independently reviewed the material and it looks sensible and the source holds up (I'll note the book is online at Google books [3]) then I don't see why the prose could not be used to start an article on the subject. --Mark viking (talk) 21:40, 18 April 2014 (UTC)

How funny, the "banner" of the banned editor edits is himself banned. Therefore, per WP:ZAUFOIHWCNA, the overwrite must be reverted. YohanN7 (talk) 21:52, 18 April 2014 (UTC)

It's just a typo. The correct link is Wikipedia:Sockpuppet investigations/Echigo mole. —David Eppstein (talk) 22:43, 18 April 2014 (UTC)

What typo? The article was turned into a redirect by Mathsci, and Mathsci is now banned. But I'm with Mark Viking: if the material by itself is found worthy to return, no matter who wrote or removed it, then it should. — HHHIPPO 09:08, 19 April 2014 (UTC)

Sorry, I misinterpreted an earlier comment. The edit summary linked to by Deltahedron contains a typo in its wikilink, making it hard to find the details of the banned user. —David Eppstein (talk) 20:22, 19 April 2014 (UTC)

I would be rather wary of restoring anything by this banned user. There was a rather bizarre case a few years ago: I cannot remember the exact details, but it involved him harassing another editor for several years by creating poor but plausible cut-and-paste articles in areas the other editor was working on, for which he was eventually banned. I had a quick look at the deleted article, and it is part of a similar pattern: it is a jumble of close paraphrases of assorted sentences from Springer's book (so in any case may be a copyright infringement). My advice is that it will probably be less trouble in the long run to rewrite the article from scratch. r.e.b. (talk) 18:50, 19 April 2014 (UTC)

Thanks for that — I'll work on a copy to eliminate those problems. Deltahedron (talk) 21:30, 21 April 2014 (UTC)

Additive polynomial and Linearised polynomial

These articles seem to overlap sufficiently that a merger could be appropriate? Deltahedron (talk) 21:15, 18 April 2014 (UTC)

User:Frivolous Consultant/sandbox

Dear mathematicians: This old Afc submission about something called "Tessellation conglomerate" will soon be deleted as a stale draft. Is there anything here that should be saved? —Anne Delong (talk) 02:19, 20 April 2014 (UTC)

There is probably room for an article on the same general topic (counting the number of copies of some shape within a larger pattern) but that draft looks pretty hopeless to me. —David Eppstein (talk) 02:33, 20 April 2014 (UTC)

Okay, thanks. I will let it go. —Anne Delong (talk) 02:57, 20 April 2014 (UTC)

Capitalization of theorem titles

I started a conversation at Wikipedia Village Pump (misc):Capitalization of theorem titles before it occurred to me it would probably be more appropriate here. I'd appreciate input. --Yoda of Borg (✉) 04:00, 24 April 2014 (UTC)

AfC submission - 25/04

Wikipedia talk:Articles for creation/Chain Decomposition. FoCuSandLeArN (talk) 15:21, 25 April 2014 (UTC)

• Wikipedia talk:Articles for creation/James Maynard. FoCuSandLeArN (talk) 20:12, 25 April 2014 (UTC)

VisualEditor math formulae

Hi everyone,

I wanted to let you know that James F is hoping to move VisualEditor's TeX-based math formula editor out of its opt-in-separately status next week. They just fixed the last batch of bugs, so it should be relatively stable at this point. This won't change the status of VisualEditor on the English Wikipedia; however, to edit math formulas, you won't have to separately opt-in to the math editor in addition to the main VisualEditor editing environment. At most other Wikipedias, it will be available to everyone.

If you haven't tried it out yet, please go to Special:Preferences#mw-prefsection-betafeatures and opt-in. If you know of (or find) any big bugs, or if there's some reason why you believe that letting everyone use this tool is a bad idea, then please let me know by the end of Wednesday (if possible). Bugs can be reported to Wikipedia:VisualEditor/Feedback if you want. Thanks, Whatamidoing (WMF) (talk) 01:44, 26 April 2014 (UTC)

Mess up formula unreadable when used with mathjax

Formula you are trying to edit is obscured and the text area for editing is way too small for complex formula

Some great big bugs.

1. When using MathJax editing a simple formula will severely mess up how its displayed. Template:Bug
2. After an editing session (i.e. clicking save) the mathjax will not be reloaded as VE does not do a full page reload. Template:BugTemplate:Bug
3. The popup box is way to small, and can obscure the text of the equation. Could be solved by making it movable and resizable.Template:Bug

These are the show stoppers for me.

The VE team should also coordinate with @Gwicke: on this issue. There are plans to completely change the maths backend. Moving away from the outdated png maths renderer to a mathml/mathjax based one with cached images. [4]--Salix alba (talk): 08:16, 26 April 2014 (UTC)

1. How long has it been since you looked at the display bug? That equation displays correctly for me in Firefox and in Safari. If it's not working for you, then please tell me your browser/skin/OS information again.
2. The page updated the formula upon save for me (and also while the math formula editor was open, after a delay of about a second or so). Whatamidoing (WMF) (talk) 00:09, 27 April 2014 (UTC)

Display bug happened just now. I even turned off all my extensions. I'm using the latest chrome on mac 10.9. I get the same with firefox. The problem with refreshing after save seems to be ok now.

BTW it does not give me a Edit Beta tab in firefox. I need to manually type the url to get to use VE in firefox. --Salix alba (talk): 06:48, 27 April 2014 (UTC)

Salix alba,

How long has this been going on? Do other things (like Twinkle) load for you? Is your screen somewhat narrow? Is the History tab displaying? Does it only happen when Wikipedia seems to be slow? If you've opted in, it should be there. Whatamidoing (WMF) (talk) 20:41, 27 April 2014 (UTC)

Also with a large formula the selection region you need to hover over to get the ∑ to appear is very small. Ideally it should be the full size of the equation. --Salix alba (talk): 11:16, 26 April 2014 (UTC)

I agree. As a workaround, you can use cursor keys to select it. Just press the arrow key until you're on top of the formula. Whatamidoing (WMF) (talk) 00:09, 27 April 2014 (UTC)

Wow, I hope that removing the TeX renderer never comes to pass. I have tried converting over to MathJax, but the MathJax typography is buggy to the point that I cannot bear to stick with it. The TeX renderer just works. --Mark viking (talk) 16:07, 26 April 2014 (UTC)

The formula editor isn't invoked for numbered equations (using template NumBlk). Also, the whole package is so slow that it tends to be confusing. It takes forever to load a big article for editing (even when you want to edit a tiny section), once you are done, it takes forever again to see the actual effects of the edits. Imo, the VE stuff should stay as beta for the foreseeable future. YohanN7 (talk) 20:23, 26 April 2014 (UTC)

The proposed change here only affects people who are already using VisualEditor. At the English Wikipedia, you would still have to opt-in to VisualEditor. Whatamidoing (WMF) (talk) 00:09, 27 April 2014 (UTC)

@Salix alba: Thanks for your thoughts – some notes and queries:

•  Template:Bug was fixed six months ago; are you still having issues with this?
• Template:Bug is a bug in the MathJax system, and does not affect VisualEditor AFAICS.
• Template:Bug is a blocked on a long-term request for improving the flexibility in the OOjs UI system on which VisualEditor is built, and won't be fixed soon.
• Template:Bug is a bug that only affects people editing multi-line formulæ who have the MathJax preference (which is marked "experimental" for a reason) switched on.
• The region being the wrong size is something we need to fix, but shouldn't be too hard.

Based on these, I think it will be OK to go ahead, and we will try to get these fixed quickly.

Jdforrester (WMF) (talk) 20:04, 29 April 2014 (UTC)

@Jdforrester (WMF): 51565 works for me now, 36060 was just for reference.

57437 has two part, movable and resizable. If you look at the above screen shot you see for a large formula the dialog window obscures what you are trying edit. If you can see the result of what your trying edit it makes for a bad ui experience.

Template:Bug is a consequence of Template:Bug which breaks any content using css-positioning. Most guides to CSS say !important is something to be avoided. For me if its a choice between having to use MathJax and the wikitext editor or VE and png/texvc thens its MathJax every time. Another minor annoyance is that with a large formula the region you need to click the mouse on is quite small, I'm having to spend time trying to find the region to click on. Template:Bug --Salix alba (talk): 21:44, 29 April 2014 (UTC)

Also bugzilla:36060 should be fixed for VE now. MathJax hooks on wikipage.content now and should redraw. The bug isn't fully fixed (because it's actually about LivePreview), where it still has a small dependency on another bug, but that should not affect VE I think. —TheDJ (talk • contribs) 22:02, 29 April 2014 (UTC)

@Salix alba and TheDJ: Yeah, the bit in 57437 about the window opening in the wrong place is actually a totally different bug ("The [selection] region being the wrong size") which we'll fix anyway. Jdforrester (WMF) (talk) 23:49, 29 April 2014 (UTC)

I think that if the problem with the too-small selection region is fixed, then the tool will automatically quit covering up the formula. Whatamidoing (WMF) (talk) 15:25, 30 April 2014 (UTC)

Is the poor performance of MathJax due to MtheJax iteslf (being a Java Script), or is it due to poor utilization of it? I am curious to know, since with modern day computers, and with today's internet technology, it's practically impossible to get something as slow as this, at least it is so if you put your minds to it. YohanN7 (talk) 22:18, 29 April 2014 (UTC)

There are multiple reasons.

1. It's not the most optimized script that has ever been written. I'm not calling it bad code or anything, but if you put a JS wizard king on it you probably can still optimize it considerably.
2. The default mode is HTML-CSS, which is the best supported cross browser. This is also it's slowest mode, but the most consistent, with the best final results (because it has the least dependencies on browser features). Note that you can choose different render modes (SVG and MathML) by right clicking a formula. MathML is suboptimal for some browsers, SVG might give very good results and is faster, if you have the proper fonts.
3. Keeping track of what works how well in which browser version is a problem in itself. MathJax for instance by default was using MathML in FF for quite a while and then disabled it again. It was a lot faster, but there were simply too many imperfections. If browser support was further along, or more consistent, MathJax could work a LOT faster.
4. Wikipedia pages are BIG. Really, our website has a lot of content, and a lot of features. It's optimized to death, but compared to more focused websites, it probably is still big. That makes something like mathjax inherently slow right now. Part of why Mathjax is so slow simply has to do with the fact that we are not able to treat every single page in the encyclopedia as a math page.

A lot of improvements can be made, for sure. The problem is that it is all dependent on a very small group of developers. I would encourage everyone to support and encourage the developers of MathJax.org. They really are doing a great job, considering the browser and operating system madness that they have to work with. Of course it is idiotic that we need MathJax to begin with, but building and proofing that MathJax CAN do it, is the best way to get this fixed in OS'es and browsers I think. Also the steps of physikerwelt and GabrielWicke referred to by Salix alba, actually prepare our 'content' for exactly these better times, while at the same time allowing us to continue to support png rendering, but adding svg rendering, semantic math (helpful for search engines) and pdf/book rendering (all problem areas right now). All enhanced with MathJax for those that desire it, or for who it makes sense. It's a long path, but at least we are seeing progress now, for the first time since 2004'ish. —TheDJ (talk • contribs) 14:37, 30 April 2014 (UTC)

Thank you @TheDJ: for the informative answer. I didn't see this until now, which explains, to some extent, my sort of (too) negative comments below. I'm all for getting MathJax rendering into good shape because of its visual appeal and the future potential of having inline math using it as well. (This is a big big no-no with the PNG rendering.) Keep up the good work! YohanN7 (talk) 09:48, 2 May 2014 (UTC)

My advice would be to try to keep down the conversation between client and server to a minimum. As I said, it is impossible to get something as slow as MathJax when running on a single machine (well, unless you use dot net of course, then everything is possible). The explanation must be that the client and server is talking to much. YohanN7 (talk)¨

• the MathJax preference (which is marked "experimental" for a reason) — really: and what reason would that be? More precisely, what are the plans for development of MathJax and its integration into VE? Just saying it's "experimental" tells us rather little. Deltahedron (talk) 06:10, 30 April 2014 (UTC)

@Deltahedron: Bugs, missing features, slowness, … – take your pick. There's been a huge amount of work by Mauritz to make MathJax a beautiful, simple, fast experience for all our users, and Jaibao, the GSoC student who worked on the initial VisualEditor plugin for formula editing last year was looking at the next step of using MathJax not only to render but also to write formulæ, as an alternative to LaTeX. I don't know if that has gone any further. Jdforrester (WMF) (talk) 17:42, 30 April 2014 (UTC)

That really is not an adequate answer to a perfectly serious question. Firstly, I don't know whether MathJax, as currently instantiated on English Wikipedia, is buggy, slow or whatever, and if I did I probably would not have asked the question. Is it any or all of those things -- I don't know and so can hardly take my pick. Secondly, "experimental" does not mean, or imply, any or all of those things. Software can be buggy or slow without being experimental, and can be experimental without being buggy, slow or deficient in features. "Experimental" implies that it is being tried out for a reason, usually with a definite set of criteria against which it will be judged, by a definite group of people and on a definite timescale. What are those in this case, please? Thirdly, I find it surprising to say the least that the Product Manager, VisualEditor team is unwilling or unable to say what the plans for mathematics rendering are or what progress has been made against those plans. Please involve the community in your thinking and I am sure we will be able to help. Give us this unserious sort of brush-off and we can not. Deltahedron (talk) 21:22, 1 May 2014 (UTC)

@Deltahedron: You seem to be confused; I have no product management relationship with the formula editor, I'm just speculating as to why it's still listed as "experimental" after several years. If you want answers, you should ask the volunteers who maintain that extension what their plans are. I don't know what they are. Jdforrester (WMF) (talk) 21:52, 2 May 2014 (UTC)

James F appears to be offline at the moment, so let me attempt a clarification on his behalf: When he says, "I have no product management relationship with the formula editor", I believe that he means "I have no product management relationship with MathJax's formula editor", not VisualEditor's formula editor. Whatamidoing (WMF) (talk) 23:01, 2 May 2014 (UTC)

Thank you for clarifying that. User:Jdforrester (WMF)'s answer seems illogical. I posted an open question on this discussion group and he answered it. He admits that his answer was merely speculation, but astonishingly tells me that it's my fault I did not get a proper answer because I should not have asked him. He then proceeds to ignore the part of the question he really ought to know the answer to ("what are the plans for [MathJax] integration into VE"). However the point of my question is:

• What plans does WMF have for mathematics rendering on en.wikipedia?
• Specifically, what are the plans for development of MathJax within WMF products and its integration into VE?

Please could we have a clear answer to those questions? As I have said before, the mathematics community here is capable of helping if only WMF would engage with us in a sensible way. Deltahedron (talk) 06:34, 3 May 2014 (UTC)

@Deltahedron: "volunteer driven".... As far as I'm aware, WMF has 0 and no plans on Math. Just as they have no plans on SVG, on 3D models, on WebP, on Timelines and graphs, on document proofreading, dictionary functionality, book writing, news writing, etc etc etc. —TheDJ (talk • contribs) 20:12, 3 May 2014 (UTC)

That may well be so. But I would like to hear a clear statement from someone at WMF in a position to make a pronouncement on this subject. It is proving surprisingly difficult to get such a thing. Deltahedron (talk) 20:20, 3 May 2014 (UTC)

@Deltahedron and TheDJ: I'm sorry for the confusion; I assumed denizens of WikiProject Mathematics, if no-where else, would be aware of the nature of engineering support in place for the mathematical functions in MediaWiki and had made a decision that they weren't going to help out those few volunteer developers working on it. Certainly it's a frequent topic of discussion when I talk to hard-science-related topic editors how the tools at the editing communities' disposal could be better supported. (TheDJ is entirely correct, as always.) Jdforrester (WMF) (talk) 05:12, 4 May 2014 (UTC)

Everyone is appreciating that good work goes into this. But this is not the problem. One problem is that performance issues must be dealt with. If not, the whole discussion will be like that when Java was entirely new. The Java enthusiasts argued that Java was bleedingly fast, when, in practice, it is was acceptably fast only if you had a super-computer able to haul the thing. I'm afraid that MathJax might turn out to be fast only if you have a 10 Giga-bit internet connection and/or a less-than-two-years-old machine. This would be ridiculous, considering that we are really only displaying text on a screen, something acceptably fast since the Windows 3.1 days. I'm still curious about why MathJax is so much slower than sever-generated PNG. Does it fetch fonts every time it runs if I don't have them locally? YohanN7 (talk) 16:53, 1 May 2014 (UTC)

Personally I don't find the performance issues with MathJax very problematic, even a heavy page like Help:Formula can be completely rendered in about 20s. I find the time of rendering roughly comparable with the time it takes to download all the png images. I do have all the MathJax fonts and the STIX fonts installed locally which make things quicker and look better. In fact the download time is always quicker with mathjax (if you have local fonts) than png. --Salix alba (talk): 23:08, 1 May 2014 (UTC)

Hmmm. Just like the old days Java arguments. It is fast (even though it is very very slow). I see the 20s or so for Help:Formula too. I have MathJax fonts installed, but not the STIX. It should take less time than a human would notice as a delay to render even that page if it was properly implemented. Granted, this is more annoying than very problematic. I tend to switch back and forth between MathJax and PNG. MathJax is much prettier, but sooner or later, I switch back because of the poor performance. YohanN7 (talk) 08:04, 2 May 2014 (UTC)

Arthur Engel (mathematician)

I created this article recently, but am really struggling with sources. I'm hoping someone here may be able to help? Thanks, Boleyn (talk) 08:01, 26 April 2014 (UTC)

It has now been referred to Wikipedia:Articles for deletion/Arthur Engel (mathematician), so any help would be appreciated. Best wishes, Boleyn (talk) 13:54, 26 April 2014 (UTC)

Merger help request

Transclusion from Requested Merger Noticeboard involving B-Class article:

• Merge Calculus and Infinitesimal calculus. Discussion >>>HERE<<<. Unsigned request -68.145.242.176 Talk 18:01, March 28, 2014‎ (UTC)

• Tagged and awaiting consensus. GenQuest ^{"Talk to Me"} 00:05, 30 March 2014 (UTC)

• Overwhelming early support for merge indicates that proposer or someone familiar with the subject(s) can proceed boldly. GenQuest ^{"Talk to Me"} 03:50, 4 April 2014 (UTC)

•  Working Consensus is for merger to proceed by interested editor or proposer. GenQuest ^{"Talk to Me"} 05:19, 16 April 2014 (UTC)

=> Can someone here help with this? Regards, GenQuest ^{"Talk to Me"} 19:37, 26 April 2014 (UTC)

Wikipedia talk:Articles for creation/Adaptive Filtering by Optimal Projection

Dear mathematicians: I'm not sure if this is the right place to report this, or if it is more physics or engineering. In any case, should this old Afc submission be kept and improved, or is there content that should be added to Adaptive filtering? —Anne Delong (talk) 02:37, 27 April 2014 (UTC)

Input requested about possible hoax

Hi all, if I could get some input at this thread, it'd be much appreciated. Regards, Daniel (talk) 13:42, 28 April 2014 (UTC)

It was a made up name with fake references now deleted Chihiro number. It is a real sequence with OEIS reference [5]. The hoax has spread to that page so if anyone has an OEIS login it would be good to fix that.--Salix alba (talk): 14:42, 28 April 2014 (UTC)

I've deleted this but the hoaxer achieved his goal of making it a DYK. It very much looks as though it was someone who knows about mathematics - changed the section heading as it would be nice to identify the hoaxer although I don't expect we will. Dougweller (talk) 14:57, 28 April 2014 (UTC)

OEIS is already on the case although it seems that their removal of the hoax link has not yet been finalized. —David Eppstein (talk) 18:04, 28 April 2014 (UTC)

I love a good hoax—and this was a good one. But I never ever said this! YohanN7 (talk) 15:18, 29 April 2014 (UTC)

Sorry I didn't do anything about this: I noticed it before it went on the front page via my occasional visits to Tomorrow's Main Page, thought it looked dubious but with my poor track record of nominating stuff for deletion which gets kept I left it alone. Should have at least posted here for more expert opinions.

I've done (created + reviewed) DYKs in the past and can easily see how this would get through. The reviewer is meant to check refs but they have to AGF for offline refs which for many articles are the only sort. There's no requirement or even recommendation the reviewer is an expert or should seek one, as long as they can understand the article enough. You can see how this was crafted to be obscure enough that it wouldn't be obviously dubious but also not so technical that it needed an expert who would be more likely to question it.--JohnBlackburne^words_deeds 18:05, 29 April 2014 (UTC)

"no requirement or even recommendation the reviewer is an expert"? Perhaps there should be? Deltahedron (talk) 18:22, 29 April 2014 (UTC)

Royal Society journals - subscription offer for one year

I'm delighted to say that the Royal Society, the UK’s National Academy for science, is offering 24 Wikipedians free access for one year to its prestigious range of scientific journals. Please note that much of the content of these journals is already freely available online, the details varying slightly between the journals – see the Royal Society Publishing webpages. For the purposes of this offer the Royal Society's journals are divided into 3 groups: Biological sciences, Physical sciences and history of science. For full details and signing-up, please see the applications page. Initial applications will close on 25 May 2014, but later applications will go on the waiting list. Wiki at Royal Society John (talk) 02:51, 29 April 2014 (UTC)

Zorn's lemma

The article on Zorn's lemma reads like a textbook, giving an example application in great detail and sketching a proof. I would like to remove these two sections to make the article more encyclopedic, but I wanted to get feedback before going forward. Brirush (talk) 16:15, 1 May 2014 (UTC)

No examples and no proof outlines doesn't automatically make anything more encyclopedic. But it might make something decidedly worse. Zorn's lemma is important enough to warrant both a proof outline and a sample application in my opinion. But the present proof and example are admittedly horribly textbooky, so I wouldn't scream out loud if you removed them. But better is to rewrite them. YohanN7 (talk) 16:34, 1 May 2014 (UTC)

I think the problem isn't so much with the content as with the writing style, which is too informal for an encyclopedia. I don't think any content should be removed, but the article should be cleaned up stylistically. Sławomir Biały (talk) 14:52, 2 May 2014 (UTC)

Just since it's so irrelevant, Wikipedia's cofounder Jimmy Wales once had a chance encounter with Max Zorn. I suspect one could find some of his comments about it via Google. Michael Hardy (talk) 14:24, 2 May 2014 (UTC)

Disambiguation help needed for Complete

Greetings! The disambiguation page Complete has 101 incoming links, putting it near the top of our list of most-linked disambiguation pages. It turns out that a lot of these links refer to one of the many mathematical senses of the word, so the expertise of this project in fixing those links would be greatly appreciated. Cheers! bd2412 T 16:32, 2 May 2014 (UTC)

Current activity

No new articles in Wikipedia:WikiProject Mathematics/Current activity for the last couple of days. Do we really have no new articles or does one of the bots need kicking again? —David Eppstein (talk) 02:22, 4 May 2014 (UTC)

Option B I think. Just one day without new articles is exceptionally unlikely. Three and something's up.--JohnBlackburne^words_deeds 02:30, 4 May 2014 (UTC)