Wikipedia talk:WikiProject Mathematics/Archive/2014/May

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 Completeness

Greetings! The disambiguation page Completeness has many 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)

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. T63497
2. After an editing session (i.e. clicking save) the mathjax will not be reloaded as VE does not do a full page reload. T53565T38060
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.T59437

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. [1]--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:

•  T53565 was fixed six months ago; are you still having issues with this?
• T38060 is a bug in the MathJax system, and does not affect VisualEditor AFAICS.
• T59437 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.
• T63497 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.

T63497 is a consequence of T54499 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. T66625 --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)

Thank you for that reply. I do not know why you would make that assumption, and as far as I personally am concerned it is quite incorrect. I have no idea what the nature of engineering support is for mathematics functions, because I have never seen any such description, I do not know where to go to find it, and my repeated questions on the subject here and elsewhere have not been answered. It is therefore impossible for me to have made any kind of decision of the sort you suggest.

However we are at least one step further on. As I understand it, User:Jdforrester (WMF), using his official WMF account has stated that the following is an entirely correct description of the WMF plans for mathematics rendering and specifically the plans for development of MathJax within WMF products and its integration into VE:

WMF has 0 and no plans on Math.

Am I the only person who thinks that the WMF's position here is lamentable? Anyway, I have (again) asked Jimbo Wales to pursue the matter with WMF [2]. Deltahedron (talk) 08:37, 4 May 2014 (UTC)

Depends, like I stated, there are a quite a few more niche areas that get 0 support. I would say undesirable, but then again, WMF can only keep in the air as many balls as it is able to juggle at that time and thus it needs to prioritize. Also remember that almost EVERYTHING we EVER did up to 2010 was volunteer driven. So why do we think everything needs to be done by WMF these days ? —TheDJ (talk • contribs) 09:49, 4 May 2014 (UTC)

Because someone has to plan and coordinate what's going on, otherwise that volunteer effort is likely to be largely wasted; because we already have complaints that the way the substantial amount of paid development work is organised is making it hard for volunteers to get their work done; because WMF is already spending a lot of time and money on radical changes such as Visual Editor and Flow which means that mathematics markup, rendering and editing is going to have to change, like it or not, and a lot of work will have to be done by someone just to sustain existing capabilities; and because WMF has raised an income of some $30M and over 100 staff precisely in order to do this sort of work. Deltahedron (talk) 10:23, 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)

Gini coefficient discussion

Project members are invited to look at Talk:Gini coefficient#Gini in Template:infobox country and to provide input. – S. Rich (talk) 04:24, 6 May 2014 (UTC)

And , if they can cope with it, Kakwani index. All the best: Rich Farmbrough, 18:22, 6 May 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)

I've removed the third proof; it was observed redundant already when it was first added, so I don't know why it survived this long. I personally am inclined to also remove the first proof and leave only the second one (which is the version that one can find e.g. in Stanley's EC2, as Thm. 7.4.4), but I will leave that for someone else to decide. I would not remove all three proofs.--JBL (talk) 15:49, 6 May 2014 (UTC)

Yes, retaining one of them is probably the right thing to do, with references, if possible, so I support your suggestion. YohanN7 (talk) 16:05, 6 May 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 [3]. 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)

Another reason for mathematicians to watch the front page and DYK queues. This slipped through:

• ... that even though the range of the Beautiful Nuthatch (pictured) is very large, approaching 376,000 km² (234,000 mi²), the species is nevertheless rare, being highly localized in its distribution?

D'oh. It doesn't especially require a mathematician, just someone with some arithmetic ability, but it got as far as the front page. I spotted it too late, just before it was replaced.--JohnBlackburne^words_deeds 18:34, 6 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)

I wrote to Jitse Niesen about this before David Eppstein's comment appeared above. Today I wrote to Oleg Alexandrov about it. Oleg runs the other bot involved. Jitse's bot edits the current activities page daily; Oleg's bot edits the list of mathematics articles daily. Jitse's bot gets its information on new articles from the edits done by Oleg's bot. Jitse's bot is still doing daily edits; Oleg's bot hasn't done its daily edits since the end of April. Michael Hardy (talk) 12:21, 5 May 2014 (UTC)

Oleg has now replied and he says he's fixed the bot. I expect all new articles created in May will appear soon. Michael Hardy (talk) 16:26, 7 May 2014 (UTC)

Probable trolling over at Naive set theory

Don't think this one is doing "good faith" edits. Seems to be just a bit too smart not to have gotten the message by now - and just about smart enough to intentionally misinterpret your attempts to reason with it in such a way as to make you the idiot.

If I revert more myself, I'll surely be banned for a century. If I post more, I'll just be feeding it. YohanN7 (talk) 11:37, 5 May 2014 (UTC)

I don't think he's a troll; I think he's a crank. Cranks inhabit a world of their own; not only can we not contact that world, more importantly, they can't contact ours. To them our logic appears to be a sequence of non-sequiturs, so there's no hope arguing with such a person. It's best to leave them alone, if possible. Since our set-theoretic friend seems intent on adding his ideas to the naive set theory page, some others of us may need to assist you in reverting him. Ozob (talk) 14:23, 5 May 2014 (UTC)

He has shown tendencies of being both a little bit of a troll and a little bit of a crank. He does understand new things (perhaps only after being allowed to figure them out for himself, making it his conclusions). The whole thing is now over at the reference desk, and is harmless. YohanN7 (talk) 14:47, 5 May 2014 (UTC)

It may be helpful to ask for page protection if an IP user persists in edit warring on an article. Deltahedron (talk) 15:31, 5 May 2014 (UTC)

Might be an idea. His last conclusion over at the reference desk was that now he had answered all the questions about why his set theory is guaranteed to be without paradoxes, and the only remaining thing was to revert the reverts of his edits in the article. If he begins, I'll ask for page protection.

Sort of funny, when you are in the age when you begin to actually learn something, but before the age when you learn more and find out that you actually know close to nothing, then you may come close to think that you know it all, almost at least. I recognize these tendencies in people, including myself in the undergraduate days. But this guy is extreme. YohanN7 (talk) 17:26, 5 May 2014 (UTC)

In light of this discussion I've semiprotected Naive set theory. If the protection turns out not to be necessary it can be lifted. EdJohnston (talk) 18:04, 5 May 2014 (UTC)

I have lifted the semi-protection since the ip-user now has a real user account. He doesn't seem prone at all to be doing pure vandalism (while not being logged in). YohanN7 (talk) 16:11, 8 May 2014 (UTC)

Where can I find the proper procedure to propose a change of the name of an article? Is it to be found in Wikipedia:WikiProposed WikiMoves or, if not, where? I am aware of the "move" tab in the article itself. The article's title is questionable, and I could see it as instead being "Informal set theory" for a few good and a few bad reasons. I'm neutral in this respect myself, and I would like to see your opinions on whether a move is a good idea. YohanN7 (talk) 17:51, 6 May 2014 (UTC)

Try Wikipedia:Requested moves. Deltahedron (talk) 18:03, 6 May 2014 (UTC)

"Naive set theory" is a term in common use within mathematics for this sort of set theory, so I think a move would be inappropriate in this case. Sławomir Biały (talk) 18:17, 6 May 2014 (UTC)

Well, except I'm not sure it's a neutral term. See the talk page of that article. I think it's at least worth entertaining the idea that a neutral article would be easier to write under a different title (for example, informal set theory). --Trovatore (talk) 22:14, 6 May 2014 (UTC)

I've added the {{Requested move}} template to the top of the existing thread at Talk:Naive set theory#Proposed move. This converts it into an official move discussion that will be listed in WP:RM and will be closed after seven days. EdJohnston (talk) 01:06, 8 May 2014 (UTC)

Leaflet For Mathematics At Wikimania 2014

Are you looking to recruit more contributors to your project?
We are offering to design and print physical paper leaflets to be distributed at WIkimania 2014 for all projects that apply.
For more information, click the link below.
Project leaflets
Adikhajuria (talk) 10:32, 9 May 2014 (UTC)

Predicate calculus

Should it redirect to Predicate logic or First-order logic? All the best: Rich Farmbrough, 18:22, 6 May 2014 (UTC).

Predicate logic, at least according to the article Predicate logic. Predicate logic includes first-order logic, but also second-order logic, etc, as well as predicate calculus (according to some), on which we don't have a separate article.

Of course, I'm just guessing, but it would make more sense to direct it to a parent than directing it to its cousin first-order logic. YohanN7 (talk) 15:40, 12 May 2014 (UTC)

Also, Predicate logic is short enough to be a disambiguation page (of sorts). The user will soon enough find a link to what he's looking for. YohanN7 (talk) 15:43, 12 May 2014 (UTC)

Cameron–Martin_theorem#Motivation

This is the most extreme case I've seen of overly abstract writing in a Wikipedia article. Everyone including most non-mathematicians has some idea how the the standard n-dimensional normal distribution behaves, but the first sentence in that section is phrased so as to be comprehensible only to analysts. I may try to rewrite it if no one beats me to it. Michael Hardy (talk) 17:47, 12 May 2014 (UTC)

Interestingly, it is so formulated from the very first version, of 2006. Boris Tsirelson (talk) 18:00, 12 May 2014 (UTC)

I think you're being a little unfair in your characterization of it. I think it could benefit from an explanation of what is meant in less abstract terms, but I don't think the statement via the Radon–Nikodym derivative is overly abstract for a topic of this sort. Sławomir Biały (talk) 18:32, 12 May 2014 (UTC)

OK, maybe I was a bit extreme, but certainly the Radon–Nikodym derivative can be written in a style that's not so notation-intensive. Michael Hardy (talk) 23:07, 12 May 2014 (UTC)

AfC submission - 13/05

Wikipedia talk:Articles for creation/A Proof of the Reconstruction of Disconnected Graphs. Is there anything salvageable there? Thanks, FoCuSandLeArN (talk) 14:44, 13 May 2014 (UTC)

No, pure OR, no encyclopedic content. --JBL (talk) 15:24, 13 May 2014 (UTC)

I agree. —David Eppstein (talk) 15:42, 13 May 2014 (UTC)

Wikidata

Just going to announce that there is now a related project on Wikidata: d:Wikidata:WikiProject Mathematics. Any help or a visit is much appreciated (signing your name and adding the page to your watchlist is even better). You may also ask questions about Wikidata's activities on the talk page. Especially input about new properties, desired data, or problems with inter-language links can be organized on that page. - And to give you an impression of what we are doing you can look at some of the data we can provide for the dodecahedron: http://tools.wmflabs.org/reasonator/?&q=178296 - Currently we can only learn that its facet is a pentagon, but hopefully much more will be available soon. --Tobias1984 (talk) 21:28, 15 May 2014 (UTC)

Firoozbakht's conjecture

Further opinions would be welcome at the article talk page. Deltahedron (talk) 09:21, 18 May 2014 (UTC)

COTM

{{Current-Math-COTM}} has been nominated for deletion -- 65.94.171.126 (talk) 07:03, 19 May 2014 (UTC)

AfC submission - 19/05

• Draft:General Purpose Analog Computer
• Draft:FlowSort. FoCuSandLeArN (talk) 22:38, 19 May 2014 (UTC)
  • The GPAC is clearly a notable topic (papers by people other than the original creator of this model hundreds of citations in Google scholar) and the article on it looks in acceptable shape. FlowSort does not appear to be notable (Google scholar finds papers mainly by the originators of the topic, not counting unrelated papers using the same name for something else, and these papers have from zero to two citations each) and the writing in the draft article is so unreadable that I think some of it must have been generated by machine translation. My suggestion would be to accept GPAC and reject FlowSort. —David Eppstein (talk) 23:00, 19 May 2014 (UTC)
    • I agree that GPAC is notable and that the article looks acceptable. I also agree that FlowSort should be rejected; and while it's a minor point, I don't think the article is the result of machine translation but is instead written in the obfuscated style of someone who has accomplished nothing but wants to appear productive. Ozob (talk) 03:37, 20 May 2014 (UTC)
      • The wonderful English phrase "mumbo jumbo" is an apt description of the latter article, in my opinion. Sławomir Biały (talk) 22:35, 20 May 2014 (UTC)
        • The creator of the second article evidently became impatient: FlowSort. --JBL (talk) 23:05, 20 May 2014 (UTC)
          • I have prodded the FlowSort article. Ozob (talk) 02:14, 21 May 2014 (UTC)

Spherical law of tangents

At Law of tangents we find this:

The law of tangents for spherical triangles was described in the 13th century by Persian mathematician Nasir al-Din al-Tusi (1201–74)

But it doesn't go on to state that law. Does anyone know it? Michael Hardy (talk) 19:28, 18 May 2014 (UTC)

Seems to be this. Deltahedron (talk) 19:33, 18 May 2014 (UTC)

The attribution may also be unclear: [4]. —Quondum 19:44, 18 May 2014 (UTC)

I've added a short new section to the article, citing the CRC book cited by Deltahedron: Law_of_tangents#Spherical_version. Michael Hardy (talk) 19:39, 22 May 2014 (UTC)

Execrable "help" with TeX offerred by web sites

There are web sites that purport to assist people in setting mathematical notation into TeX or MathJax code. So I am told by people who use them, but I have no URLs to cite. They write things like this:

{ {f} \left({ {{{x}}} + {{{ y }}} }\right)^{{ {2} }} } + { \frac { expression } { { { expression } } } } - { \sum_i { {a}_{i} } }

It would suffice to write this:

f(x+y)^2 + \frac{expression}{expression} - \sum_i a_i

Obviously this can make editing difficult and one failure to count superfluous left and right curly braces while editing can destroy a line, and the error can be hard to find.

I've seen this a number of times in the past few months, and I just did this edit. My purpose in doing that edit was only to break a line into two lines using "align", but I cleaned up some of this sort of crap while doing it.

Is there some way to avoid having this pestilence infect all Wikipedia articles? Michael Hardy (talk) 20:18, 22 May 2014 (UTC)

For what it's worth, in many cases probably the actual source of things like this is machine-generated LaTeX code. --JBL (talk) 20:25, 22 May 2014 (UTC)

Ask the people who are in charge of the programs that generate that to fix them. Getting rid of superfluous brackets from expressions has been a reasonably popular task for undergraduates in computing since year dot. Dmcq (talk) 09:24, 23 May 2014 (UTC)

Half-exponential

The new article titled Half-exponential function has just two articles linking to it: a "see also" link from Exponential function and an appearance in the list of exponential topics. The article is about compositional square roots ƒ of exponential functions, i.e.

$f(f(x))=e^{x}.\,$

The article mentions growth rates, so maybe some growth-rate- and Landau notation-related articles should link to it.

So:

• Which other things should link to it?
• What else should be done with it?

And should we have an article titled compositional square root? Michael Hardy (talk) 19:02, 22 May 2014 (UTC)

There is Functional square root, partially in answer to all three questions. Sławomir Biały (talk) 19:32, 22 May 2014 (UTC)

OK, now I've redirected compositional square root to functional square root. Michael Hardy (talk) 19:40, 22 May 2014 (UTC)

How do we know that such a function even exists? If we do not, then perhaps this article should be deleted. Or changed to say that such a function cannot exist. JRSpriggs (talk) 05:56, 23 May 2014 (UTC)

The article functional square root refers to a paper by Kneser in which the functional square root of the exponential function is constructed. That source should obviously be consulted before we entertain the idea of deletion. Sławomir Biały (talk) 10:33, 23 May 2014 (UTC)

It's been known for some time that the functional square root of any increasing function f satisfying x < f(x) or x > f(x) exists, and it can be taken to be C^k or C^∞ if the original function is. Supposedly, there is an real analytic solution, but I haven't verified that. — Arthur Rubin (talk) 19:51, 23 May 2014 (UTC)

These comments about C^k functions should be in the functional square root article. — Arthur Rubin (talk) 20:06, 23 May 2014 (UTC)

Leaflet For Wikiproject Mathematics At Wikimania 2014

Are you looking to recruit more contributors to your project?
We are offering to design and print physical paper leaflets to be distributed at Wikimania 2014 for all projects that apply.
For more information, click the link below.
Project leaflets
Adikhajuria (talk) 13:46, 25 May 2014 (UTC)

Is anyone interested in taking this up? We would need to provide the following:

Your contact details	If we need any clarification about your project - won't go in the leaflet.
Project name
Short Description	50 words. Must be intelligible to non-Wikimedians.
Longer description	150 words. Why should they help? How can they get involved? Should include key links.
Logo	(image from commons, ideally in svg, if not at least 800px width by 500px height)
Primary webpage URL
Mailing List
Email contact	Email for primary contact/community manager/comms
IRC channel
Facebook page
Twitter handle

Does this project have a logo? Is File:Glass tesseract still.png usable? What about the social media? Deltahedron (talk) 19:41, 26 May 2014 (UTC)

I'm not wild about the glass tesseract image. My preference would be for something having more of a fractal flair, like File:3-adic integers with dual colorings.svg. Sławomir Biały (talk) 22:37, 26 May 2014 (UTC)

More general definition of line and surface integrals?

The articles line integral and surface integral base their definitions on a (smooth) parametrization of the curve and surface respectively. But my textbook defines the following instead:

1. Let f be a scalar function of three variables. The surface integral $\iint _{S}f\ dS$ over a surface S is defined by $\lim _{\lVert P\rVert \to 0}\sum f\Delta S$.
2. Let g be a scalar function of two or three variables. The line integral $\int _{C}g\ ds$ along a curve C with respect to arc length is defined by $\lim _{\lVert P\rVert \to 0}\sum g\Delta s$ where s is the arc length of C.
3. Let h be a scalar function of two or three variables. The line integral $\int _{C}h\ dx$ along a curve C with respect to x is defined by $\lim _{\lVert P\rVert \to 0}\sum h\Delta x$. Likewise with the other spatial variables.

In all cases, the definition includes "provided the limit exists", and ||P|| denotes the norm of the partition of a Riemann sum. In the case of the surface integral, the partitioning is done rectangularly and the norm ||P|| is the length of the longest diagonal of a rectangle among all the partitions $\Delta S$. For the line integrals, the norm of the partition is defined analogously to the one-dimensional Riemann integral. In both cases, the sum is of all partitions over C or S. For vector fields, the definition is component-wise (e.g. the definition of work can be split (in three-dimensional space for a steady force field) into three integrals of the third kind above). Unlike the line integral's labeling of case 2 a line integral of a scalar field and case 3 a (component of) a line integral of a vector field, all three definitions can be applied to both vector and scalar fields.

I think these are better than the current definitions because they are more general with no conditions set on the integration domain - only that the limit has to exist. The book also gives the definitions in the articles as "Evaluation Theorem"s, but attaches the additional conditions such as C having to be piecewise smooth. The only concern I have is whether the expressions I typed above are well-defined. To me, it would seem they are well-defined because of the given definition of ||P||.--Jasper Deng (talk) 23:33, 26 May 2014 (UTC)

I wonder, which textbook do you read. I remember a well-known counterexample to "The surface integral $\iint _{S}f\ dS$ over a surface S is defined by $\lim _{\lVert P\rVert \to 0}\sum f\Delta S$." Namely, the limit does not exist for a cylinder (and, even, a constant function on it). The problem is that a small size of a rectangle does not ensure that the normal vector to the rectangle is close (in direction) to the normal vector to the cylinder. Though, really, it was about triangles, not rectangles. But I am still in doubt. What is called "rectangular" in this context? Is it proved that the limit exists at least in some nontrivial (that is, non-flat) cases? Boris Tsirelson (talk) 05:41, 27 May 2014 (UTC)

The "rectangular" partitioning is really referring to parallelogram-shaped pieces of a curved surface. To elaborate on how the book makes this well-defined, each parallelogram is a subset of the tangent plane to the surface at each particular point chosen for each $f\ dS$. That a tangent plane exists is because the book does indeed assume that S is smooth (or at least "piecewise smooth", i.e. a finite union of smooth surfaces like the cylinder; in that case we sum over each smooth portion and add them, just like the case with a piecewise smooth curve with line integrals). This definition is independent of whether the surface is orientable or not. But yes, the smoothness of S goes to imply that normal vectors exist. Nevertheless, I still maintain that this definition is more general than the one given in the article. One advantage is that S doesn't have to have a regular projection onto any of the planes.--Jasper Deng (talk) 06:07, 27 May 2014 (UTC)

I am pretty sure that it is not so simple. I think Spivak presents the cylinder counterexample in his Calculus on Manifolds. It might also be that Federer discusses this in his Geometric Measure Theory (a quite advanced but very beautiful book), but I don't recall for certain.

In addition, I'm not so sure that it's necessary for the line integral and surface integral articles to have extremely general definitions. Because the article will be consulted on a regular basis by physicists and engineers, and because piecewise smooth sets are both much simpler than the general case and are the most common application, I think it's best if the article start with the piecewise smooth case and only mention generalizations later. Ozob (talk) 14:17, 27 May 2014 (UTC)

I would like to strongly endorse Ozob's last sentence: the articles on surface and line integrals should aim be readable by anyone whose background consists of a standard semester of computational multivariable calculus. --JBL (talk) 15:26, 27 May 2014 (UTC)

Unindenting since I don't know where else to put this. At least locally it is true that for a regular surface

$\int f\,dS=\lim _{\|P\|\to 0}\sum f(x_{ij})\Delta S_{ij}$

where the partitions are determined relative to some fixed parametrization of the surface, and $\Delta S_{ij}$ is the surface area of the ij-th part. This does not contradict the Schwarz paradox since one is not approximating the surface by flat pieces; only the scalar function f is being approximated by its value at some points in the partition. The analysis is needed to show that the areas ΔS_ij are equal to the value of the Jacobian plus little oh of the mesh size, in order to obtain the formula $dS=|r_{u}\times r_{v}|du\,dv$ that our article surface area gives. Sławomir Biały (talk) 20:06, 27 May 2014 (UTC)

Yes, this should be correct. But what is gained this way? The surface should be smooth enough. True, the function need not be continuous, rather it needs to be Riemann integrable... but this is the same as Riemann integrability in the parameters, if we parametrize the surface... nothing new happens. Boris Tsirelson (talk) 20:33, 27 May 2014 (UTC)

Nothing in particular is gained. I do not mean to suggest that this is necessarily a good approach, but I just wish to point out that it does not lead to contradictions as has been suggested here. It might be marginally preferable to do things this way when intuition is more important than rigorous proofs. (Why else would we use the Riemann integral in the first place?) The formal mathematical approach is actually rather intricate, invoking both partitions of unity and change of variables. At the end of the formal process of defining the integral in a way that is both well-defined and diffeomorphism invariant, it is actually difficult to validate our physical or geometrical intuition. It fails to give a satisfactory answer why the resulting integral should be identified with the surface integral. (One good approach to answer the latter question that can be done in a serious course on the subject is to replace the surface integral by the integral over a tubular neighborhood, thus "fattening out" the surface, and then pass to a limit.) But there's probably no royal road. Sławomir Biały (talk) 21:29, 27 May 2014 (UTC)

Yes, this was kinda what I was concerned about the quality of my book's definition, which deliberately is restricted to surfaces where tangent planes exist everywhere (requiring differentiability). Essentially, I'm wondering about how the notion of a surface element $\Delta S$can be well-defined without using the unbent approximation using tangent plane segments.--Jasper Deng (talk) 01:47, 28 May 2014 (UTC)

What you are really asking for is a measure. For n-dimensional objects in n-dimensional Euclidean space, there is only one reasonable definition, Lebesgue measure (at least when doing geometry; for probability there are all kinds of useful measures). For smooth (or even C¹) m-dimensional objects in n-dimensional Euclidean space, again, there is only one reasonable definition. This is essentially because one can locally choose a differentiable parametrization of the set and pull back the measure to the domain of the parametrization; the change-of-variables formula for integrals says that the only reasonable thing that one can get (again, assuming that we are doing geometry) is Lebesgue measure on m-dimensional space times the absolute value of the Jacobian of the parametrization. Rougher sets, however, cannot be measured in this way, because when there is no C¹-parametrization there is no change-of-variables formula. One has to invent a new method for measuring these sets, and it turns out that there is more than one way to do so. Probably the best known is Hausdorff measure, but there are others. Handling such sets consistently is a subtle thing, but by now the technology is well-developed. The full story is too long to get into here, but the essential content is that you pick some sets, decide what their sizes ought to be, and use them to approximate other sets. Any textbook on measure theory will explain how to do this; in particular, the book of Federer that I mentioned above carries this out in the geometric situations you're interested in (though it sounds like it's rather more advanced than what you're reading). Textbooks on measure-theoretic probability also do this (but motivated by other considerations coming from stochastic processes). Ozob (talk) 04:27, 28 May 2014 (UTC)

Really, I agree with Slawomir and Jasper: something essential is gained. Boris Tsirelson (talk) 05:55, 28 May 2014 (UTC)

Robert Steinberg

An anonymous IP posted to his article that he had died. Can anyone confirm this? Deltahedron (talk) 06:02, 27 May 2014 (UTC)

An obvious place to look for such information is at his institution:

In Memoriam: Robert Steinberg

Posted on May 27 2014, 12:03 pm

Bob was born on May 25, 1922 in Soroki, Bessarabia, Romania (present day Soroca, Moldavia) and came to settle in Canada with his parents when he was still very young. He was a student of Richard Brauer in Toronto, receiving his Ph. D in 1948 before joining UCLA in 1948 where he stayed till the end. . .

This suggests the news is true :( Rschwieb (talk) 13:25, 28 May 2014 (UTC)

Thanks for that. It was indeed an obvious place to look: so obvious that I had already checked but when I did so the announcement had not yet been posted [5]. (The time stamps make that, well, obvious.) Deltahedron (talk) 20:38, 29 May 2014 (UTC)

List of higher derivatives of the inverse function

I think it would be useful to have an article containing a list of the higher derivatives of the inverse w.r.t. composition of a function from the reals to the reals. We already have Inverse functions and differentiation#Higher derivatives, but that only goes to the third derivative and only shows the derivatives in Leibniz's notation. I think that it would be desirable to go to much higher derivatives (say at least the tenth) and to also show the derivatives in Lagrange's notation.

Does such an article already exist? If not, should we not create it? JRSpriggs (talk) 18:58, 30 May 2014 (UTC)

These higher derivative are rather straightforward to compute with a computer algebra system, either for a specific function or for a general function. The output of such a computation is much more useful than a list in WP, which can hardly be used without copy errors. Therefore, such a list would have a much lower encyclopedic content than an explanation of how one may construct this list with, say, Maple and Mathematica (this is not difficult, but needs some expertise of these software). D.Lazard (talk) 19:34, 30 May 2014 (UTC)

Are there suitable reliable sources to support such an article? Deltahedron (talk) 19:48, 30 May 2014 (UTC)

Not everyone has access to computer algebra systems. Mathematica can output TeX. Why not have it produce the list in Wikipedia-ready format?--agr (talk) 19:56, 30 May 2014 (UTC)

Relationship between mathematics and physics

Relationship between mathematics and physics is a new article that could use work. Michael Hardy (talk) 15:14, 29 May 2014 (UTC)

That article seems very essay-like and unencyclopedic. I don't know what its fate should be. Sławomir Biały (talk) 16:37, 29 May 2014 (UTC)

The article has plenty of sources that demonstrate the subject is notable, so I think it's mainly a matter of improving the article. It has only existed for two days - we should help the editor instead of hemorrhaging tags. RockMagnetist (talk) 17:26, 29 May 2014 (UTC)

I didn't say the topic wasn't notable. It would be nice if someone would rewrite the article based on secondary sources. At present, it seems rather like a novel synthesis of primary sources. Sławomir Biały (talk) 17:34, 29 May 2014 (UTC)

Sorry, I misunderstood. Usually when someone talks about an article's fate, it's thumb up or down. To me, it seems like a historical survey (with a touch of cheerleading), and it needs more discussion of the philosophical problems. RockMagnetist (talk) 17:43, 29 May 2014 (UTC)

I agree that the article isn't well developed at this point. The vague title begs the question, what should it become? There is already a discussion of the philosophical relation between math and physics (at least from a physicist's point of view) in The Unreasonable Effectiveness of Mathematics in the Natural Sciences and Quasi-empiricism in mathematics. Should the article be more about the historical interplay between mathematical and physical developments? Or should it wax poetic about physical aesthetics and the beauty of mathematically elegant physical theories? (The last is in jest, but is often brought up in popular accounts of math and physics.) --Mark viking (talk) 20:18, 30 May 2014 (UTC)