Wikipedia talk:WikiProject Mathematics/Archive/2015/Jul

Jacobian conjecture
One or more math editors are needed to deal with suggestions I posted to the talk: Jacobian conjecture page. If you need to contact me, please use email, as I have difficulties navigating to my user talk page. — Preceding unsigned comment added by L.Andrew Campbell (talk • contribs) 00:40, 3 July 2015 (UTC)

uncorrelated random variables
I just moved uncorrelated to uncorrelated random variables. Normally that would cause the new title to appear in the new articles list tomorrow. But it won't unless someone finally rescues Jitse' bot. Michael Hardy (talk) 19:07, 3 July 2015 (UTC)

AfC submission
Is Draft:Hierarchical testing of variables in high dimensional datasets notable? Best, FoCuSandLeArN (talk) 20:30, 3 July 2015 (UTC)

Need a math person with some hard copy texts
I am trying to track down and alleged copyright violation. A reader wrote to the Wikimedia foundation noting that the text of the following article: Cyclic decomposition theorem is a copy of the source material:


 * Hoffman, Kenneth; Kunze, Ray (1971). Linear algebra (2nd ed.). Englewood Cliffs, N.J.: Prentice-Hall, Inc. p. 231

Our usual copyright tools failed to identify a problem because they tend to work with online materials. I tried invoking one tool which has access to off-line academic material but that failed to identify a problem. I don't believe the reader is making it up but I need some confirmation before I can remove it. I'm hoping that there is some reader of this wiki project who happens to have a copy of the text handy and can check it out.-- S Philbrick (Talk)  15:05, 4 July 2015 (UTC)


 * The theorem statement is a word-for-word copy of the one appearing in the source. The editor who added the material in question is correct that inevitable there will be some overlap in any two statements of the same theorem, but an exact match cannot have been a mere coincidence.  I believe that this does qualify as a copyvio from the rather stringent conditions of Wikipedia.  The article should be deleted, and replaced with an untainted version.   Sławomir Biały  (talk) 15:41, 4 July 2015 (UTC)
 * Thanks for the response, now deleted. (and a candidate for a new article, if someone is looking for something to do :)-- S Philbrick (Talk)  17:16, 4 July 2015 (UTC)

Is Jitse's bot dead forever?
Jitse's bot hasn't updated the new articles list in almost two weeks. Four days ago I sent Jitse Niesen an email and I haven't heard anything. Will Wikipedia be forever dependent on one person in that way? How should we proceed? Michael Hardy (talk) 23:27, 26 June 2015 (UTC)


 * Actually I responded at the very top section several days ago and got no response. I will try again here. What does it take to run the bot? Does it run on the wikimedia foundation's server? I found this page  but I'm not sure if is related. As for a long-term plan, I agree it would make sense if someone else can take over running the bot (it should be so much easier to taking over the task than starting over), as apparently Jitse is no longer active. -- Taku (talk) 00:17, 29 June 2015 (UTC)
 * I've raised it here: Bot_owners' noticeboard.-- JohnBlackburne wordsdeeds 18:55, 3 July 2015 (UTC)

I think Jitse Niesen will probably look at this within 48 hours. Michael Hardy (talk) 03:59, 9 July 2015 (UTC)

Jacobian conjecture Please Help
I really need some help from experienced math editors. I posted some suggestions to the talk:Jacobian conjecture page on June 23, designed to correct some basic mathematical flaws in the article. I am a genuine expert on the subject but I am now blind and totally inexperienced with Wikipedia. So others need to carry out the actual edits. The suggestions don't correct everything in the article, but make a good start. Someone must know how to bring this to the attention of previous editors of the page.Please use email to contact me if desired.

L.Andrew Campbell (talk) 00:53, 9 July 2015 (UTC)


 * It looks to me like some of those suggestions have already been implemented by User:R.e.b. several days before this post. In fact, unless I'm mistaken, all of your suggestions were promptly implemented.  I'm more than happy to help, but could you please be clearer what still needs to be done?   Sławomir Biały  (talk) 13:26, 9 July 2015 (UTC)


 * Sławomir Biały, I think that L.Andrew Campbell probably has difficulty reading this page. I was thinking of sending your message to him as an e-mail, but I would only do that if you haven't done so already.  Have you?  --JBL (talk) 14:48, 9 July 2015 (UTC)


 * No, I haven't.  Sławomir Biały  (talk) 22:42, 9 July 2015 (UTC)
 * Just saw those changes, thanks. I will try to make further changes myself.

L.Andrew Campbell (talk) 23:31, 9 July 2015 (UTC)

Draft:Thaine's theorem - references available if someone knowledgeable can apply them
Dear mathematicians: There appear to be quite a few book references that could be added to make this stub into an article, for example,  and , but I don't know how to apply them. Can someone help? &mdash;Anne Delong (talk) 01:28, 9 July 2015 (UTC)


 * Why not simply add these to the stub and push it out into main space? If someone is interested, then the article will expand in the natural Wiki fashion.   Sławomir Biały  (talk) 13:42, 9 July 2015 (UTC)
 * , It seems that has done just that.  Thanks to you both. &mdash;Anne Delong (talk) 13:25, 10 July 2015 (UTC)

E (mathematical constant)
Two separate editors have requested some outside input at Talk:E (mathematical constant). See Talk:E (mathematical constant) concerning a question of whether we should cite someone's self-published and unpublished computer programs (including a rather novel interpretation of WP:SPS), and Talk:E (mathematical constant) concerning whether original research is permitted. Sławomir Biały (talk) 12:55, 9 July 2015 (UTC)
 * Other Wikipedia math experts please comment. Rather than continue discussion on the talk page he sites above, Slawomir took it upon himself on the same day he posted the comment above (9 July 2015) to delete much of the table of "Known Digits" at E (mathematical_constant) and substitute "Since that time, the proliferation of modern high-speed desktop computers has made it possible for amateurs to compute billions of digits of e". Please look at what he deleted. He is totally missing the documented and referenced and very important algorithmic advances that made those increased approximations possible. His page change comment is "WP:SPS is pretty clear that we shouldn't have these entries..." when that was the very thing being discussed with me at Talk:E (mathematical constant). He just decided he was right and stopped the discussion by deleting much of the table. He also says "There are no recognized authorities for world record computations [of e], unlike pi" without siting what those pi authorities are or why the history of pi precision at Pi is included in Wikipedia but the page on e somehow doesn't warrant such information. Also in Talk:E (mathematical constant) he comments "Anyone with a computer can compute billions of digits of e, as the case of Alexander Yee shows." Again, he is minimizing and missing the algorithmic advances by mathematicians and computer scientists (such as myself) who spent significant time and effort advancing the state of the art using extended-precision approximations of e as an example problem. Someone with the authority, please correct him and undo his changes or explain the inconsistencies between the pi and e articles in such a way that makes it obvious why such different standards should be applied to the two. Thank you. Rick314 (talk) 04:43, 10 July 2015 (UTC)
 * From the article "The number of known digits of e has increased substantially during the last decades. This is due both to the increased performance of computers and to algorithmic improvements." Two references are given.  If you feel that details of the algorithmic improvements have been omitted, you are welcome to supply additional references to reliable secondary sources.
 * The reason there is a difference between &pi; and e here is that there are many reliable secondary sources concerning the quest for more digits of pi. For example, Arndt and Haenel spend a chapter discussing it.  Papers in "Mathematics of computation" and "The American Mathematics Monthly" discuss new algorithms for computing &pi;.  Indeed, Borwein and Borwein have and entire book on this subject. As far as I am aware, no such sources exists concerning the computation of e, although I have requested such sources, as they would certainly be most welcome if they existed.  On the contrary, the portion of the table I removed was sourced to usenet posts and unpublished emails.  Those are simply not acceptable sources, at either article.  So there aren't "different standards ... applied to the two", rather I have applied the same strict standards of sourcing to the one article as the other.
 * In fact, I have given the e article slightly the benefit of the doubt by including two references to Gourdon concerning "algorithmic advances". Since you believe that I have "missed the algorithmic advances", now would be a good time to correct me by supplying sources to the secondary literature (peer reviewed journals, books, etc.) that discuss these advances.
 * Finally, regarding the rather childish observation that you made "He just decided he was right and stopped the discussion by deleting much of the table", in this post, you agreed that "Existing references already violate [our guidelines]." I agreed with this statement, and removed the portion of the table that did not agree with our guidelines.  My edit summary clearly referenced the guideline that they violate, and I also referred in that discussion to policy reasons why Wikipedia should not be in the business of record-keeping on the authority of Usenet posts and unpublished emails.  I would have thought that was clear.  There was evident consensus that these entries did not meet the requirements set forth at those policies and guidelines.  Your arguments did nothing to contradict that conclusion.  Indeed, one of your arguments is "Look at my documentation referenced in this discussion above. It is actually better than several already existing table references and I have emails that confirm third-party (not me) verification of my results."  Yes, that's true.  But the other results were poorly-referenced as well.  Surely it is not very surprising that the outcome would be their removal, rather than the inclusion of your own results in the table, given our apparent agreement that the references used in the table are unsuitable?   Sławomir Biały  (talk) 13:37, 10 July 2015 (UTC)
 * I agree with Rick314 that it was rather poor form for Sławomir Biały to go make this edit in the middle of a discussion. But on substance I think Sławomir Biały is correct, for exactly the reasons he has given. --JBL (talk) 13:53, 10 July 2015 (UTC)
 * Well, Rick314 had already announced his intention to make an edit, against consensus, guidelines and policies. There was a very clear consensus at the discussion page that the table entries were inappropriate.  The discussion was ongoing for several weeks, during which time anyone at all could have weighed in.  If we waited indefinitely for such discussions to be "done", nothing would ever happen on Wikipedia, and we would be stuck in endless discussions with COI/SPAs.  So I disagree very strongly that it was,"poor form".  This is just an articulation of WP:BOLD, which is policy.   Sławomir Biały  (talk) 15:04, 10 July 2015 (UTC)
 * Sorry, but this is a ridiculous number of words and policies to spout on a question of basic politeness and consensus-building. If you're discussing changes with someone and haven't reached consensus, it's rude to go and impose your version while discussion is ongoing.  There was no particular rush, and there had been no third opinion yet.  You acted a bit like a jerk, unnecessarily.  The world would have been a very slightly better place if you hadn't done so.  That is all. --JBL (talk) 15:14, 10 July 2015 (UTC)
 * Excuse me, but I really think thus criticism of my behavior is way out of line, and I think you should apologize. As far as I can tell, I'm the only editor at E (mathematical constant) who is urging against a series of ill-considered edits that go against policoes. I feel that the community has been negligent here.  I shouldn't be the only one needing to comment there.  I know this is a well-watched page.  So where was everybody else?  How come I was the only one holding the bag there?  Twould weeks went by, and made an edit, with clear reasons behind it. By any standards, those are reasonable actions.  And calling other people names also isn't likely to make the world a better place.  Sławomir Biały  (talk) 15:56, 10 July 2015 (UTC)
 * Notice after Slawomir's edit the article reads "...the proliferation of modern high-speed desktop computers has made it possible for amateurs to compute billions of digits of e." There is no mention of algorithmic improvements there. It refers to Alexander Yee as an example amateur. (He is not.) Similarly, amateur me improved expert Steve Wozniaks's algorithm 4 ways as explained on the first 2 pages of my 1992 source code. So no apology from me, and both Slawomir's process and content are inappropriate. "Expert" and "amateur" are subjective terms to be avoided, and I think intentionally used in the article by Slawomir as an insult. Let's stick to Wikipedia publishing guidelines. As I read WP:SPS and WP:ABOUTSELF (look at the details) the original table entries seem fine. Regarding books, does e: The Story of a Number (Princeton Science Library) move e and pi closer together in terms of leaving a history of precision improvements? The Table of Contents is available at that link, but I am not sure what content makes it more or less applicable to this discussion.  — Preceding unsigned comment added by Rick314 (talk • contribs) 17:22, 10 July 2015 (UTC)
 * There is no mention of algorithmic improvements there. This is wrong.  From the article:  "The number of known digits of e has increased substantially during the last decades. This is due both to the increased performance of computers and to algorithmic improvements."  I said as much above, and requested additional citations.  References to the literature are welcome.  Tendentious lies like this are not constructive, and could potentially result in having your editing privileges revoked.
 * Regarding "amateurs", Yee's own website refers to this as a high school project. It's hard to defend your view that he was acting in a professional capacity as a mathematician.
 * We already reference "e - the story of a number". And, in fact, according to that source, which contains almost no discussion of records, "In one sense the stories of e and &pi; differ.  Because of the longer history and greater fame of &pi;, the urge to compute it to an ever greater number of digits over the years has become something of a race... No such craze befell e."  So, if anything, this source actually supports the opposite stance.
 * You're just wrong about WP:SPS. That policy states "self-published media, such as books, patents, newsletters, personal websites, open wikis, personal or group blogs (as distinguished from newsblogs, above), Internet forum postings, and tweets, are largely not acceptable as sources."   The weird interpretation of the policy that it's ok to use self-published sources in scientific articles as long as one is discussing one's own accomplishments falls well outside of any reasonable reading of it, especially taken in conjunction with the rest of our policies.
 * Finally, you have a clear conflict of interests insisting that the article must cite your own unpublished work. This makes it essentially pointless to attempt discussion, since you are clearly using a very self-serving reading of policies, guidelines, discussions, and the article.  For example, you apparently cannot be bothered to notice that the article does mention algorithmic improvements, even after it has been pointed out to you!
 * The bottom line is, we are absolutely not going to cite your usenet posts and unpublished work. If you have a paper that is published in a peer reviewed place (typical of professional, as opposed to amateur, mathematicians), then please give a reference to that.  Otherwise, I don't see further discussion as remotely constructive.   Sławomir Biały  (talk) 18:03, 10 July 2015 (UTC)


 * My initial reaction was the same as 's but I in no way disagree with 's edit. Coming to the discussion late and reading the whole thing at once gave the impression that the edit was rash, but I now realize that the time element (this discussion had been going on for weeks) was missing. I was about to make the same edit and upon finding out that it had already been done my initial reaction was just re-enforced. Perhaps it would have been better if another editor had done it, but as Sławomir has pointed out, we (collectively) were late to the game and dropped the ball on this one. As to the substance of 's objections, I find no merit in the argument and see this as falling squarely under the umbrella of WP:OR. If Rick314 publishes his results and an unrelated third party writes about them then we can include a discussion of them by sourcing the third party. Although many do chaff under these restrictions, they are necessary to prevent WP from devolving into yet another internet bulletin board. There are many breaches of these guidelines to be found in our articles, and when we come across them we need to act decisively, as Sławomir has done. Bill Cherowitzo (talk) 18:22, 10 July 2015 (UTC)
 * Thank you Bill. You said "If Rick314 publishes his results and an unrelated third party writes about them then we can include a discussion of them by sourcing the third party." I asked about this way way back in the article Talk. My dated algorithm documentation and source code and execution metrics and more are now published in Wikimedia Commons. I can put them somewhere else if appropriate. I have emails with a third party and they checked everything there. My reviewer could have entered a row for my 1988 and 1992 records? (Slawomir please stop repeating your position and I will too. I hope we can get input from others and the amount of text here is unwieldy.) Rick314 (talk) 01:23, 11 July 2015 (UTC)
 * I think you misunderstood something. In this context, "publishes" means to publish in a way that makes something a reliable source, e.g. by going through peer review and getting it published in an academic journal. It doesn't mean merely making something available online. —David Eppstein (talk) 03:04, 11 July 2015 (UTC)
 * Sorry that I did not make myself clear. When I say published, I mean published in a peer reviewed venue and not any form of self publication and the same holds for the third party source. Notice that no one here has criticized your work or even hinted that it might be incorrect, in fact I am working under the assumption that it is perfectly fine – the objections have to do with how statements that appear on WP are backed up "in the literature". The editors of WP are not to be the judges of the content, they can only report on what the reliable secondary sources say about that content. If there are no such sources, then we do not say anything and don't include the content! Bill Cherowitzo (talk) 03:17, 11 July 2015 (UTC)
 * OK, and thanks for all your help and explanations. I better understand Wikipedia content rules now. Rick314 (talk) 18:10, 11 July 2015 (UTC)

Correlation coefficient
Can someone convert this disambiguation page into an article explaining the general concept of a correlation coefficient, and how all the types on the page fall into this concept? Cheers! bd2412 T 14:13, 13 July 2015 (UTC)


 * Does correlation and dependence help ? Gandalf61 (talk) 14:57, 13 July 2015 (UTC)
 * It's already a link on the page. Correct me if I'm wrong, but I presume that a correlation coefficient is merely a coefficient that illustrates correlation (in the correlation and dependence sense)? bd2412  T 15:25, 13 July 2015 (UTC)
 * Yes, that's right - it is a quantitative measure of correlation (of some type). Gandalf61 (talk) 15:43, 13 July 2015 (UTC)
 * Ok, that will do. Thanks! bd2412  T 16:12, 13 July 2015 (UTC)

Rayleigh II
I disambiguated "Rayleigh II" at Wave field synthesis to point to Rayleigh–Jeans law, but now I'm not so sure if that is correct. I find references to a Rayleigh I, Rayleigh II, and  Rayleigh III equations (See, e.g., Ahmet Kondoz, ‎Tasos Dagiuklas, Novel 3D Media Technologies (2014), p. 213, noting the use of the Rayleigh I and Rayleigh II equations for plotting wave scattering), which fit topically but differ in the equations used. What is the right answer here? bd2412 T 17:37, 15 July 2015 (UTC)

OR in progress at Thomas precession
Folks, those with an interest in relativity, please take a look at what I believe is unjustifiable, unsourced OR in progress at Thomas precession. Look at [this claim], for example. —Quondum 17:39, 12 July 2015 (UTC)


 * Isn't this is an overreaction ? YohanN7 and myself have been creating diagrams for Thomas precession which do appear in a few papers by at least two authors. We are filling in the gaps where these papers are unclear, so by WP:CALC it should be allowed. M&and;Ŝc2ħεИτlk 18:08, 12 July 2015 (UTC)


 * For clarity... YohanN7 has, as always, provided painfully detailed explanations and Quondum walked off, because "bully boy" YohanN7 was being too "energetic".


 * When the diagrams were created, Quondum provided no source, nor reference, nor even his own worded description, supporting his claim that Lorentz boosts in any direction do not preserve orthogonality of the spatial axes, he just accused YohanN7 of "unsourced" WP:OR and took the report here. I blindly (as always) assumed I was wrong and others are correct, and thought I knew where Quondum came from and only blundered, about the LT matrix for a boost in any direction (no longer on Talk:Thomas precession, see its history if you care). Quondum has also claimed we were "not giving room to speak", which is clearly false because he has and no-one stopped him.


 * Another reference supporting orthogonality under Lorentz transformations (rotation and boost) is Goldstein, 2nd edition, page 285. M&and;Ŝc2ħεИτlk 16:34, 16 July 2015 (UTC)

Reply


Since I have been reported for allegedly doing original research on a talk page, I feel I have to be my own lawyer and defend myself. I do thank Maschen for some moral support, here, and there.

First off, everything is in order because Quondum's claim's about unsourced original research are false. Two of the sources are presented below. Normally Cushing should more than suffice, but since weight is needed versus authority, I added Weinberg as well, almost notorious for not making mistakes, both here and in the article.

To the right there is also an excerpt from Cushing's paper that verbatim verifies the disputed passage and diagrams in the article, and also my post on the talk page for which I am being reported.

At least in theory, it is thus a possibility that Quondum is incorrect in real life. But this matters little. He is most certainly wiki-incorrect, because my statement is verifiable. According to Wikipedia policy, verifiability weighs more than truth.

Quondum's allegation that it is unjustifiable is possibly true. We can, as is done in most (actually all) of the relevant literature, toss the whole issue off as being obvious. That would save valuable space in the article. But, alas, Quondum wouldn't have that to begin with, so, we have to go through this in depth in the article.



Original research
So that wasn't original research, whether unsourced or unjustifiable. But this is! Let


 * $$\mathfrak{so}(3,1) = \mathfrak{so}(3) \subset \mathfrak{so}(3,1) \oplus \mathfrak k,$$

vector space direct sum. Let $ξ ∈ k$ be an arbitrary boost parameter. Then $e^{ξ ⋅ K}$ is, by definition, an arbitrary pure boost. Claim: For any rotation parameter $ς$,


 * $$T = e^{-\boldsymbol \zeta \cdot \mathbf J}e^{\boldsymbol \xi \cdot \mathbf K}e^{\boldsymbol \zeta \cdot \mathbf J}

= e^{e^{-\boldsymbol \zeta \cdot \mathbf J}(\boldsymbol \xi \cdot \mathbf K)e^{\boldsymbol \zeta \cdot \mathbf J}}$$

is a pure boost (equality is by property of the matrix exponential). By a property of $Ad$,


 * $$e^{-\boldsymbol \zeta \cdot \mathbf J}(\boldsymbol \xi \cdot \mathbf K)e^{\boldsymbol \zeta \cdot \mathbf J} \in \mathfrak{so}(3,1).$$

By the relationship between $Ad$ and $ad$,


 * $$e^{-\boldsymbol \zeta \cdot \mathbf J}(\boldsymbol \xi \cdot \mathbf K)e^{\boldsymbol \zeta \cdot \mathbf J}

= e^{\mathrm{ad}_{-\boldsymbol \zeta \cdot \mathbf J}}\boldsymbol \xi \cdot \mathbf K.$$

But


 * $$[-\boldsymbol \zeta \cdot \mathbf J, \boldsymbol \chi \cdot \mathbf K] \in \mathbf k,$$

for arbitrary $χ$, so (recalling $ad_{Z}^{n}(W)$ works by composition)


 * $$e^{-\boldsymbol \zeta \cdot \mathbf J}(\boldsymbol \xi \cdot \mathbf K)e^{\boldsymbol \zeta \cdot \mathbf J}

= e^{\mathrm{ad}_{-\boldsymbol \zeta \cdot \mathbf J}}\boldsymbol \xi \cdot \mathbf K \in \mathbf k.$$

Hence


 * $$T = e^{e^{-\boldsymbol \zeta \cdot \mathbf J}(\boldsymbol \xi \cdot \mathbf K)e^{\boldsymbol \zeta \cdot \mathbf J}}

= e^{-\boldsymbol \zeta \cdot \mathbf J}e^{\boldsymbol \xi \cdot \mathbf K}e^{\boldsymbol \zeta \cdot \mathbf J}$$

is a pure boost, again by definition. Since $exp$ is onto both $SO(3)$ and $SO^{+}(3,1)$, and since the action of $SO(3)$ on a sphere is transitive, the claim in the by Quondum disputed [edit] is proved (once the remaining easy details left to the reader are filled in).

If one follows the action of the successive transformations on the spatial coordinate axes, it is apparent what is going on. Rotate, then stretch the $z$-axis, Oops, this is wrong, the boost is active leaving all coordinate axes alone boost (leaving coordinate axes intact), then rotate back.

The interpretation is not unsourced original research, see Cushing above and A. Ben-Menahem (1985) (ref in article) for stretching by $γ$ For strikeout, see above small comment. Since references are scanty on proofs, I might, by WP:CALC, put this in a footnote in the article, unless someone spots an error. YohanN7 (talk) 17:03, 13 July 2015 (UTC)

Tricki
I was just looking around Tricki at and thinking how it serves a purpose almost completely orthogonal to Wikipedia in helping with how things are done. I was wondering if there wasn't some way of getting a bit of synergy somehow where we continue along the current paths but people can easily find a way over to Tricki from the reference desk for instance. It doesn't fit in with Wikipedia's content policies and I can't see an easy way of doing anything with it but it seems a shame not to have a way of guiding people who ask questions to a place that helps them with how to solve a problem. Dmcq (talk) 10:20, 15 July 2015 (UTC)
 * Probably a first step would be to have a WP article on it other than just a redirect to Timothy Gowers. Maybe a second step would be to jut start linking WP articles to Tricki where appropriate. For reference desk questions there's no reason you can't supplement an answer with a link to Tricki if it would help; there are a couple of occasions where a question was answered with a link to math.stackexchange.com or mathoverflow.net. I browsed an article, on Gowers's weblog on the future of Tricki and he was saying that he started Tricki before these sites had taken off, and they have taken away a lot of the potential online math market. But to extend your analogy, I think the space has enough dimensions for quite a few mutually orthogonal directions. Another relatively recent development is the appearance of online lectures and screencasts. A role WP could play in this is as a kind of switchboard to these other resources. I'm not certain that the "External links" format we have now is up to the task though. When WP first started there were maybe a few other websites on specific topics, but nowadays you can take an entire lecture course (sans final exam), find a selection of textbooks for it and even get help with questions, all online and often free. --RDBury (talk) 17:45, 16 July 2015 (UTC)

Sine
Sine is undergoing an AfD suggesting that it should be merged back into Trigonometric_functions. Please participate at Articles for deletion/Sine. —David Eppstein (talk) 00:35, 18 July 2015 (UTC)

Also, there's a fairly new deletion sorting page, WikiProject Deletion sorting/Mathematics, that you might find it useful to watch and/or check regularly, since that's where things like this should eventually appear. —David Eppstein (talk) 07:39, 18 July 2015 (UTC)

Request for Comment
Comments would be appreciated on this RfC: Talk:Infinite_monkey_theorem. Thanks DaveApter (talk) 10:29, 20 July 2015 (UTC)

Independence and its negation

 * (Z \perp\!\!\!\perp Y)_{G_{\overline{X}}} \qquad(Z \not{\perp\!\!\!\perp} X)_G
 * $$(Z \perp\!\!\!\perp Y)_{G_{\overline{X}}} \qquad(Z \not{\perp\!\!\!\perp} X)_G $$

How can we get this negation to look better? Michael Hardy (talk) 00:38, 19 July 2015 (UTC)
 * Even more negative space?
 * (Z \perp\!\!\!\perp Y)_{G_{\overline{X}}} \qquad(Z \not\!\!{\perp\!\!\!\perp} X)_G
 * $$(Z \perp\!\!\!\perp Y)_{G_{\overline{X}}} \qquad(Z \not\!\!{\perp\!\!\!\perp} X)_G $$
 * (As discussed here.) I would suggest doing it in wikiformatting using the double up tack unicode character instead (⫫) but getting that to work well with a double-subscript-overbar is unlikely, not to mention that amazingly unicode seems to be missing the negated double up tack symbol. —David Eppstein (talk) 01:28, 19 July 2015 (UTC)
 * Is such symbol in use (anywhere except, maybe, wikipedia)? I never saw it (as well as negated orthogonality, etc). Moreover, my experience with some economists shows that they tend to misunderstood the meaning of "not independent", mostly as "either positively dependent or negatively dependent" or something like this. The exact meaning of this negation seems to be clear only to mathematicians; but these do not need such notion (I think so). Boris Tsirelson (talk) 04:25, 19 July 2015 (UTC)
 * I've seen this symbol used for probabilistic independence used in other places. Michael Hardy (talk) 21:19, 20 July 2015 (UTC)
 * That double up tack unicode character doesn't display for us with old computers. YohanN7 (talk) 11:29, 19 July 2015 (UTC)
 * I would do it with a few more slash-bangs, as David suggests. Also, I would probably use the built-in symbol \amalg ($$\amalg$$) instead of hacking TeX to produce such a symbol out of other ones.   Sławomir Biały  (talk) 12:22, 19 July 2015 (UTC)
 * \not\amalg ($$\not\amalg$$) actually works without spacing hacks, but unfortunately that's the wrong symbol: conditional independence uses the double up tack. —David Eppstein (talk) 16:12, 19 July 2015 (UTC)

I've implemented David Eppstein's solution at Instrumental variable. Michael Hardy (talk) 21:18, 20 July 2015 (UTC)

Ratios always dimensionless?
Pls weigh in: Talk:Ratio. Fgnievinski (talk) 05:49, 22 July 2015 (UTC)

Perennial upright d
The perennial upright d issue has been raised once more at WT:MOSMATH. Opinions are welcome. Sławomir Biały (talk) 17:02, 24 July 2015 (UTC)

Jacobian conjecture results
I have posted suggestive changes with the above title to the JC talk page. An experienced math editor is needed. As a wrinkle, they include use of the name Druzkowski, which needs a dot over the letter z. Thanks,

L.Andrew Campbell (talk) 23:03, 24 July 2015 (UTC)

Anomalous MathJax rendering.
In Variance, some code said this:
 * \sigma_y^2 \le 2y_\text{max} (A - H)

I changed it to this:
 * \sigma_y^2 \le 2y_\max (A - H)

Strangely, this had the effect of causing BOTH "max" and the left parenthesis to appear in subscript. So I changed it to this:
 * \sigma_y^2 \le 2y_{\max} (A - H)

That's how it stands now. Let's try it both ways and see how it looks here:
 * $$ \sigma_y^2 \le 2y_\max (A - H), $$
 * $$ \sigma_y^2 \le 2y_{\max} (A - H), $$

Just a bug, or is there some sensible reason for this? Michael Hardy (talk) 20:25, 24 July 2015 (UTC)
 * I think the basic principle is that some or most of our math rendering pathways are not actually based on Knuth's TeX code and have occasional differences in parsing because of it, that can usually be fixed with more braces. —David Eppstein (talk) 20:57, 24 July 2015 (UTC)
 * I've use \max as a subscript with no braces in Wikipedia articles thousandds of times over quite a few years without seeing this problem. Will it suddenly be necessary to change all of them because of some change that just happened in the software? Michael Hardy (talk) 07:11, 25 July 2015 (UTC)
 * It seems like almost correct behaviour. \max is a maths operator which expects the next character as an argument. ( is the next character so that is taken as the argument. Running the the code through standard LaTeX produces an error and won't compile at all. It does not look like its new behaviour as you get the lower bracket in PNG, and MathJax. --Salix alba (talk): 08:59, 25 July 2015 (UTC)
 * It looks like the problem only happens when the next character is a bracket, so most case are OK. I've run a search for equations with the bug and fixed the ten or so pages where it occurs.--Salix alba (talk): 11:20, 25 July 2015 (UTC)
 * Would it be possible to find every case where \max occurs in a subscript or superscript to change it to {\max}, maybe even without needing to do each one by hand? And \min too. Michael Hardy (talk) 19:48, 25 July 2015 (UTC)

Uniform convergence
A new user Sergey Liflandsky is adding proofs to the "Uniform convergence". Neither mathematics nor English are free of errors. But more important is the question, do we need these proofs at all? Boris Tsirelson (talk) 11:00, 26 July 2015 (UTC)


 * I don't think the proofs are an improvement. Generally speaking, it's acceptable for Wikipedia articles to summarize proofs, but not to give detailed proofs of statements.  In the article under discussion, the proof is already summarized as the &epsilon;/3 trick.  Filling in the details is routine.  To readers for whom it is not routine, then WP:NOTTEXTBOOK applies.  Detailed proofs are the sort of thing one looks for in textbooks, not encyclopedia articles.   Sławomir Biały  (talk) 11:19, 26 July 2015 (UTC)


 * I see the same editor has added a proof to Weierstrass M-test. This is a little more justifiable I think, because there was a proof already there and the new proof is somewhat more straightforward than the old one.  Ideally, both proofs should be shortened and summarized.  In fact, the new proof can more or less be reduced to a sentence or two.  Something like: "Because the sequence of functions is uniformly bounded in absolute value by a summable sequence of numbers $$M_n$$, the difference between two partial sums of the series of functions is bounded by the difference between two partial sums of the series $$\sum M_n$$.  So the Cauchy criterion implies that the series of functions converges."   Sławomir Biały  (talk) 11:31, 26 July 2015 (UTC)


 * Also Ratio test. Boris Tsirelson (talk) 11:52, 26 July 2015 (UTC)


 * I don't object to that one. The test is rather baffling without a proof.   Sławomir Biały  (talk) 13:11, 26 July 2015 (UTC)


 * I got rid of the proof that Riemann integration commutes with uniform limits. This is geometrically obvious, and the given proof was very unenlightening.   Sławomir Biały  (talk) 12:38, 26 July 2015 (UTC)


 * It is also possible that some proofs are notable by themselves. I even wrote an article once because I thought the proof of something (at least in one direction) was extremely nifty. YohanN7 (talk) 14:52, 27 July 2015 (UTC)

Naming convention for Parenthetical disambiguation
At hypercycle (geometry) which I renamed to hypercycle (hyperbolic geometry)  there is been some discussion about "Parenthetical disambiguation" and we would like to have that discussion in a more general place (here).

The facts (add when needed) :


 * There is a need to have a parenthetical disambiguation:
 * There are two meanings of Hypercycle, one hypercycle (chemistry) and a mathematical one that is only used in hyperbolic geometry neither one should be the primary subject for Hypercycle


 * wp:DAB: If there are several possible choices for parenthetical disambiguation, use the same disambiguating phrase already commonly used for other topics within the same class and context, if any. Otherwise, choose whichever is simpler. (no more specific guidance exist)

So what is the best parenthetical disambiguated name for this page? The options are (add when there are more):


 * hypercycle (mathematics)
 * hypercycle (geometry)
 * hypercycle (hyperbolic geometry)

Why hypercycle (geometry)?
 * Historical, the article was named this way previously
 * Shortest parenthetical disambiguation name

Why hypercycle (hyperbolic geometry)?


 * The subject (as described in the article ) is rather sopecialised and only used in hyperbolic geometry and I think the most precise parenthetical disambiguation is best. WillemienH (talk) 05:57, 27 July 2015 (UTC)

Why hypercycle (mathematics)?
 * (no reasons yet), but is this not what wp:DAB prescribes?

Please join in WillemienH (talk) 05:57, 27 July 2015 (UTC) (ps I prefer hypercycle (hyperbolic geometry) )
 * Your choice of "(hyperbolic geometry)" violates the "whichever is simpler" part of the guidance you quote, so I don't think it is a good choice. I don't have a strong preference between the other two, but I tend to prefer "(geometry)", in part because we already have plenty of geometry disambiguators and in part because it was the one that was already used. —David Eppstein (talk) 06:39, 27 July 2015 (UTC)


 * I agree with David and others that "(hyperbolic geometry)" is unnecessary complicate, as it does not help readers in their choice. On the other hand, my opinion is that "(mathematics)" must also be avoided in this case. In fact, I am not sure that "hypercycle" is not used in other parts of mathematics. In any case, as "cycle" is used with various meanings, in particular in graph theory and topological algebra, it is useful to inform the reader that "hypercycle" is not a generalization of such cycles. Therefore my opinion is that the move done by must be reverted. D.Lazard (talk) 09:25, 27 July 2015 (UTC)
 * There are certainly papers that study differential equations or dynamical systems intended to model the chemical kind of hypercycle. There are also papers about hypergraphs that call certain structures in them hypercycles. So yes, "mathematics" seems too ambiguous and this gives another reason to stick with "geometry". —David Eppstein (talk) 23:46, 27 July 2015 (UTC)
 * My understanding is that the purpose of parenthetical disambiguation is primarily for typing into the search box. As such, shorter is almost always better.  I think hypercycle (geometry) is best.   Sławomir Biały  (talk) 12:14, 27 July 2015 (UTC)

Help on disambiguating "matrix models" needed
Hello, math people! Me and i think most other wikipedians who are focusing on eliminating ambiguous links in Wikipedia are stumped by a number of math-y articles that link to Matrix model disambiguation page. Please see this "dablink-list" (which shows 7 articles now), hit "FIX" on any one of them, and see if you can identify which version of "matrix model" should be linked. If you don't know, just exit / delete the tab that has opened up. It is easy and fun to use this cool disambiguation-related tool, but you have to know something about the content area to fix these seven ones. TIA, -- do ncr  am  07:03, 29 July 2015 (UTC)

Branches of geometry
Any opinions of this new article? Michael Hardy (talk) 00:09, 29 July 2015 (UTC)
 * It does seem a bit less unwieldy than list of geometry topics, but... Is there some sourcing for this structure, or is this just some editor's opinion of what the important topic groupings are? Why projective and affine geometry but not inversive geometry? Why is Minkowski geometry listed as a subset of affine geometry, and why is it included at all when we have no article specifically on that topic? Is the first group of topics really classified by axiomatization, rather than by symmetry groups (per the Erlangen program)? Why is Riemannian geometry listed as a top-level topic instead of being a subtopic of differential geometry? Where are convex geometry, discrete geometry, computational geometry, and synthetic geometry supposed to fit? Is this a listing of branches of geometry, or a listing of geometrical branches of modern mathematics research, sweeping the parts of geometry that get taught to schoolchildren under the single Euclidean geometry link? —David Eppstein (talk) 00:24, 29 July 2015 (UTC)


 * [edit conflict] I don't know where to begin. For one thing, such classifications are usually original research, in my experience. Mgnbar (talk) 00:28, 29 July 2015 (UTC)
 * I may add to the points set by David that the relation between the different branches of geometry are much more complex that that suggested by the structure of this article. For example, projective geometry and affine geometry are both considered in synthetic geometry as well as in analytic geometry, and none; of these branches are uniquely defined by sets of axioms. IMO, this is in geometry that the main branches of geometry, and their relations should be described. Therefore, I suggest to merge this article in geometry. D.Lazard (talk) 08:34, 29 July 2015 (UTC)


 * I started that page, on this page my idea was to construct some kind of scheme that sorts out the parent (less axioms) -child (more axioms) relations between the different (mostly synthetic geometrys, but haven't found a lot of time for this yet (and also became confused, to name two:
 * There are different opinions what non-Euclidean geometry means: Is it absolute geometry or is spherical geometry also a part of it.
 * Is Minkowski geometry really an affine geometry? (Not all axioms of incidence geometry in particular every two points are on a line)
 * The only reference I have till now is
 * Maybe it is not the right page for a page about this, but I think such a page would be an good addition to WP WillemienH (talk) 09:44, 29 July 2015 (UTC)


 * The list is also missing global analysis/geometric analysis, pseudo-Riemannian geometry, non-commutative geometry, elliptical geometry, tropical geometry. Also, one might look at the AMS classification to see other areas that we've missed.   S ławomir  Biały  13:32, 29 July 2015 (UTC)
 * This first thing that stood out for me was the division between geometry with different axioms vs other types. You can divide geometry by synthetic vs. analytic, and also by Euclidean vs. hyperbolic vs projective vs. whatever, but these are orthogonal classifications. Projective geometry can be treated synthetically or analytically just as easily as Euclidean and the same goes for many other types. So while I'm not sure that the division in the article is wrong, the way the topics have been split between them seems arbitrary. Also, it's not clear how this article, when the missing branches are filled in, would differ meaningfully from List of geometry topics if some of the clutter was removed from that article. In other words it seems like we now have two articles with the same purpose. --RDBury (talk) 21:21, 29 July 2015 (UTC)