Wikipedia talk:WikiProject Mathematics/Archive 68

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Borel functional calculus

From what I can tell, the article Borel functional calculus seems to be riddled with misleading mis-representations (see talk page); if anyone feels expert on this, and have the urge, please fix it up. Right now, it seems to pretend its holomorphic functional calculus, or something(?), followed by vague hand-waving; I can't get anything out of it. linas (talk) 04:06, 30 November 2010 (UTC)

You could actually be wrong about this? Charles Matthews (talk) 08:05, 30 November 2010 (UTC)
Charles, there are multiple faults in this article, and in measurable function. I think I documented all of those failings quite clearly. The combination of all of those errors and failings left me very confused, which provoked a lot of hot air, but no fixes to any of the articles. Telling me that I am 'wrong' does not in any way make the articles clearer, or more accurate, or less misleading. Please try to be helpful. linas (talk) 03:44, 1 December 2010 (UTC)
The article was authored by User:CSTAR, who knows what he is talking about in this area. Your comments on the Talk page there show at least one confusion. It has been explained to you, there. If you can point to a mathematical mistake, then of course I'd be concerned to fix it. Charles Matthews (talk) 15:32, 1 December 2010 (UTC)

Jan Łukasiewicz has no picture

There is no picture of Łukasiewicz in our article, and none available on commons. This is an unfortunate omission. Łukasiewicz lived 1878–1956 so there should be something available somewhere, either in the public domain or by fair use. —Mark Dominus (talk) 16:47, 30 November 2010 (UTC)

Hmm. Most online sources seem to reuse variants of this image of Łukasiewicz. Interestingly, the "copyright information" there states that they believe (but do not positively know) it to be in public domain. I didn't find anywhere any clue where the image originally comes from, to verify its copyright status.—Emil J. 18:09, 30 November 2010 (UTC)

Nomination for deletion of Template:In

Ambox warning pn.svgTemplate:In has been nominated for deletion. You are invited to comment on the discussion at the template's entry on the Templates for discussion page. Thank you. (talk) 05:26, 2 December 2010 (UTC)

Template: Optimization algorithms

I removed a lot of heuristics from the template on optimization algorithms. I removed heuristics that IMHO are not discussed in optimization textbooks and journals (regardless of their merits in IEEE transactions on swarming fireflies, etc.). Second opinions would be desirable. Thanks! Sincerely, Kiefer.Wolfowitz (talk) 11:16, 30 November 2010 (UTC)

I have nothing against the heuristics, but the template was getting pretty bloated so I think this is an improvement. CRGreathouse (t | c) 06:15, 1 December 2010 (UTC)
I removed the line on "statistical methods" since they are not discussed in optimization textbooks (like the minor heuristics I removed earlier). I created a one-line entry on combinatorial optimization, using the items from the MIT book on algorithms (Rivest, Cormen, etc.). I removed many individual quasi-Newton methods, leaving only BFGS and L-BFGS (despite Lancelot's giving a comeback to SR1). Kiefer.Wolfowitz (talk) 11:51, 2 December 2010 (UTC)

Unreferenced articles

How is this project doing on referencing all its articles (at least once)? I have just found one that was easily referenced by searching Is there a significant backlog? Charles Matthews (talk) 05:28, 2 December 2010 (UTC)

Short answer: yes. Long answer: Wikipedia:Pages needing attention/Mathematics/Lists#Articles lacking sources (and the several sections following that one, also about badly sourced articles). —David Eppstein (talk) 05:42, 2 December 2010 (UTC)

Thanks. There are certainly some ones that are easy to do at the EoM still there. People might trust us on quadratic equation, and even sine wave (now has a reference). But I realised after doing a few that I'm supposed to remove items from that list by hand? Old technology, it would seem.

In the spirit of the first item in Wikipedia:Wikipedia Signpost/2010-11-29/News and notes, could we have an effort to trim that list down? I much prefer maintenance categories as a way to look for such things; but in any case some automation would help; and removing the standard material from that list would also be a good way to find lurking less-standard articles. Charles Matthews (talk) 06:55, 2 December 2010 (UTC)

A while ago I took a stab at doing the really old ones, i.e. unreferenced since 2006. There's a project that's going through all the articles in chronological order, but the more specialized articles, like some of the math ones one, really need someone who understands the material to endure the reference being added is actually related to the subject.--RDBury (talk) 09:31, 2 December 2010 (UTC)

My point is that some do appear to be really easy to get done. Charles Matthews (talk) 11:44, 2 December 2010 (UTC)

ƒ or f?

Hello guys, it seems that the consensus on Wikipedia talk:WikiProject Statistics is that f is preferred to ƒ. Does the same go for articles under the jurisdiction of WikiProject Mathematics? I expressed my objections to the consensus there as follows:

I strongly disagree with you guys. ƒ is clearly better than f. Almost all mathematicians these days write in LaTeX, and their math-mode f looks almost identical to ƒ. Take a look: f,\ g,\ h (compare that to ƒ, g, h) . Tradition dictates that the f's be more fancy than the g's and h's, which I think out-wieghs the need for consistent fonts. Ideally, the math-mode feature on Wikipedia would be more consistent with the rest of the article, so we could write all mathematical symbols on WIkipedia in math-mode, but that's not the case. As such, using the ƒ, g, and h convention is the best way to denote functions.

Just want to know.--Dark Charles 21:55, 19 November 2010 (UTC)

I think ƒ (ƒ) is better because it is how Latex typesets f in math mode, and it more clearly sets f apart from the surrounding text. Many experienced editors at WPM use this. Also, the accessibility problem mentioned a Wikiproject Statistics seems to be a non-issue. I have never encountered a browser that could not display the character. (Although I suppose it's possible that some peculiar screen readers might read it as "florin", that's an issue that we can address if it is actually the case rather than just hypothetically.) Sławomir Biały (talk) 22:39, 19 November 2010 (UTC)
My problem with "ƒ" is that the only way I know how to generate it is by cutting it from a text which already contains it and pasting it where I want it. How do you make it? JRSpriggs (talk) 00:32, 20 November 2010 (UTC)
It should perhaps be in the "math and logic" section of the character-insertion script under the edit window. But it produces the same result as ƒ anyway. Algebraist 00:41, 20 November 2010 (UTC)
Sławomir, LaTeX does not typeset f in math mode using a special character—it is simply a regular lower-case letter f in a typical italic serif font. The lower-case letter f in italic serif fonts usually extends below the baseline, as in this sentence. When LaTeX is using sans-serif fonts for math, such as in slides produced with the popular Beamer package, it is clear that the character used is an ordinary italic letter f, not some special hooked version; see [1] for example. —Bkell (talk) 00:45, 20 November 2010 (UTC)
Umm... f\,\! looks just like ƒ to me. Whether you want to call this a "special character" (your words not mine) is irrelevant semantic quibbling. Also, the f displayed in your sentence looks nothing like the f as it displays in Latex. Also, I don't see why a link to someone's beamer slide is relevant to this discussion. Obviously, when I said that Latex typeset things this way, I meant that Wikipedia's implementation of Latex did so. Sławomir Biały (talk) 01:55, 20 November 2010 (UTC)
I don't think it's very useful to go looking for Unicodes that look a little more like what renders in LaTeX. The LaTeX looks different because it's a different typeface, not a different character set.
What we really need to do is get Wikimedia to provide us a way of rendering LaTeX that doesn't look like crap. The current PNG renderings are just appalling, especially when you consider that LaTeX in general is usually the way of producing the best-looking documents (there may be a way to get results of comparable quality with Word, but I have no idea how). Blogs on seem to do really really well with rendering LaTeX; maybe we could do something like that? --Trovatore (talk) 02:02, 20 November 2010 (UTC)
mathscinet and mathoverflow use mathjax. Is there a good reason to oppose using this? RobHar (talk) 02:16, 20 November 2010 (UTC)
If it means anything, I checked Principles of Mathematical Analysis by Rudin and he uses the fancy f (i.e. ƒ), which is meaningful because the book is written in Times New Roman. However, I checked my version of LaTeX, and the italics f and the math mode f were the same... Nonetheless, I still think ƒ is better because (1) it looks more like the TeX formatting in Wikipedia and (2) it looks like the f's for functions that I grew up looking at in math textbooks.--Dark Charles 03:04, 20 November 2010 (UTC)
You know that while you are worrying about f versus ƒ, wikipedia is replete with a used to denote a? I'm with Trovatore, the global solution improving the latex display is what is actually needed here; presumably this will happen soon enough (with some work), so effort put towards f versus ƒ is probably going to be wasted in the end. RobHar (talk) 03:17, 20 November 2010 (UTC)
I can agree with that. If there is actually going to be a serious push to improve math rendering in Wikipedia, and something is really going to be done about it, then the debate about f and ƒ will be moot. —Bkell (talk) 05:11, 20 November 2010 (UTC)
I really don't care much about f vs ƒ. The important question is rendering math in general. The mathjax samples look great; I don't know what the technical issues are or how well it works in general, but if it's as good as the samples and can be done, we should absolutely do it. --Trovatore (talk) 03:07, 20 November 2010 (UTC)
The rendering from mathajax looks much better that what we currently have, but I don't see how using it would affect f vs. ƒ; unless Wikipedia changes the default font to match that used for math (or—shudder—vice versa), the f in running text still won't match that in rendered mathematical material. I don't agree with Sławomir that it's “irrelevant semantic quibbling”; were that the case, we'd treat 1 and l, and a minus sign and en dash as interchangeable. The bottom line is that choosing characters based solely on a appearance is a bad idea; among other things, what looks fine with the default typeface can look terrible if a reader has set preferences to use a different typeface. JeffConrad (talk) 06:11, 20 November 2010 (UTC)
The point Bkell was making above is that the f in Latex is just an ordinary italic f in serif font, which was blatantly not so as even his own example shows. If there is a serif font that produces an ƒ as its f, I guess that would work and everybody should be happy, since it isn't a so-called "special character". At any rate, the symbol ƒ is a "function symbol", as mandated by the unicode standard (see [2]). So I think its purpose here is clearly consistent with its intended use, much as the minus sign from unicode is distinguished from an endash. So, hypothetical user preferences that make the function symbol look like something else would have to be seriously broken indeed! Sławomir Biały (talk) 14:36, 20 November 2010 (UTC)
Sławomir, do none of these italic serif f's extend below the baseline for you? Times New Roman: f; Times: f; Garamond: f; Bodoni: f; Century Schoolbook: f; Caslon: f; Georgia: f; Palatino: f; Antiqua: f; generic serif: f. Now you might not have all of those fonts installed on your computer, but hopefully you have at least one of them, and all of those that you have installed should extend below the baseline. It is very, very common for serif typefaces to produce something that looks like ƒ for a lower-case f in the italic font. If none of these examples here work for you, just find a book or newspaper printed in a serif font and find a lower-case f in ordinary text. —Bkell (talk) 17:21, 20 November 2010 (UTC)
The Times and Palatino look ok, although there seems to be a real danger that these will not display very consistently. Are you proposing one in particular be used as a replacement for the "special character"? I think the point about unicode providing this symbol for this very purpose is not one which should be dismissed lightly though. The character is part of the standard unicode character set, not requiring any additional special fonts to be installed. Sławomir Biały (talk) 17:43, 20 November 2010 (UTC)
No, I'm proposing no such thing. All I'm saying is that the f used in math where a serif typeface is being used is just an ordinary lower-case f in the corresponding italic serif font. It is not a variant shape used for math—it is the same shape as an italic f in the serif typeface, the same shape as an italic f in regular text. In particular, compare the output of LaTeX using the default Computer Modern typeface: \,\textit{ordinary\ serif\ italic\ text}; \,f(x)=3x^2. Note that the letter f in math is the same as the ordinary text italic f. That's what I've been saying all along. Therefore, if the text is set in a sans-serif typeface, the proper shape for a math f is the same as an italic (technically, oblique) letter f in the sans-serif font, not a variant character such as ƒ. —Bkell (talk) 22:22, 20 November 2010 (UTC)
It may *look* the same as an f in a different font, but it is not the same character. It is a different unicode position. Representing the unicode ƒ by changing fonts and using an f is a hack, and an unnecessary one at that, because the unicode character itself works perfectly well. —David Eppstein (talk) 22:29, 20 November 2010 (UTC)
Well, LaTeX doesn't use Unicode, so I'm not sure what you mean. The f\, in the math italic font used by LaTeX has the same character code as the letter f in the ordinary text italic font and in the roman font, so I don't see how you can claim it is a different character without claiming that all of the math italic letters are different from the corresponding text italic letters. If you are claiming that professionally typeset math uses different characters for "math f" and "text italic f," I'm going to have to ask you for some evidence to support that. Please understand that I am certainly not advocating the ad hoc usage of a font change for functions called f. (Ignore my comment in which I posted a bunch of samples from various fonts; I was just doing that for Sławomir, because apparently my example sentence from earlier wasn't showing up in an italic serif font for him, for otherwise he would have seen that serif italic f's generally extend below the baseline. It was not a proposal of any kind.) —Bkell (talk) 22:46, 20 November 2010 (UTC)
So, to clarify what I'm trying to say: I understand that f and ƒ are two different Unicode characters. I also understand that ƒ is described by the Unicode standard as, among other things, "function symbol." But my claim is that professionally typeset math, such as that produced by LaTeX, does not make this distinction—a "math f," in professionally typeset math, is the same character as a "text italic f," just as a "math g" is the same as a "text italic g." —Bkell (talk) 23:28, 20 November 2010 (UTC)
I think the idea is that if we can improve the display of math so that it meshes seamlessly with the running text, then all math, including that in running text, can be enclosed within <math> tags, producing a uniform appearance. —Bkell (talk) 06:16, 20 November 2010 (UTC)
That's certainly one approach, and the one I initially used, but the more I looked at it, the more apparent the glaring clash between the mathematical symbols and the running text became. Consequently, I now do all mathematical symbols in the text using ordinary HTML and Wiki markup. I'd suggest looking at recent ISO standards, in which the running text is sans serif but all quantity symbols are serif; to my eye, it looks as though it was set by a blind monkey. I'd actually rather set displayed mathematical material in the default typeface than mix faces without a mighty good reason. The folks who designed type for hundreds of years before we had computer or even typewriter thingies were not idiots, and in most cases, it would behoove us to study their example and learn from it. Something else to ponder: what would Don Knuth do? In any event, this is somewhat of a side issue; I think we'd be far better off switching to mathajax, and of course using the proper logical characters rather than resorting to typographical malapropism. JeffConrad (talk) 08:26, 20 November 2010 (UTC)

A while back, I did some testing and was able to get jsmath to work with Wikipedia (mostly). So I have complete confidence that someone with more skill than me could get mathjax working. The difficulty would be convincing the developers that it's worth their time to add it. When I contacted them about the issue, the opinion I detected was "the current system is good enough". So it would be an uphill battle to get Wikipedia converted over to mathjax. — Carl (CBM · talk) 13:28, 20 November 2010 (UTC)

I think that there is fairly clear consensus that the current system is not good enough. This discussion became perennial a long time ago. Sławomir Biały (talk) 14:01, 20 November 2010 (UTC)
I agree with Trovatore and anyone who would go for a fundamental solution to this and other problems. Issues like this thread and also the one concerning the standalone formulas discussed above and numerous other discussions of the same gist in the history of WP:MATH show that the current suboptimal rendering is binding a whole lot of resources. Maybe we as the ones most suffering from this should develop an agenda what steps should be taken to implement any better solution? Does anyone know what the concrete steps would be to implement mathjax on MediaWiki? I guess we should speak up to whoever developer(s) is in charge as a WikiProject.
Another question, maybe stupid: this example on mathoverflow does not look like this mathjax integrates much more smoothly into the surrounding text. Could this be overcome just by choosing another font? I figure changing the font for the usual (and maybe Greek) letters will be not a big deal, but I wonder what changes would come with this, e.g. would size and shape of \int have to change, too, etc.? Jakob.scholbach (talk) 14:04, 20 November 2010 (UTC)
Apparently, there's already a mediawiki page describing how to install mathjax on mediawiki software [3].
Also, re font size, it looks like mathjax can scale its output, so everything would be shrunk at once, both the letters and symbols, which is presumably what we would want (see here). RobHar (talk) 14:44, 20 November 2010 (UTC)
PS: The University of Toronto math department uses mathjax on its mediawiki software, so if you browse around you'll see some examples of how mathjax could show up on wikipedia ( RobHar (talk) 14:58, 20 November 2010 (UTC)

Maybe someone should start an RfC that we can present to the devs to show that there is a clear consensus that a fundamental solution such as MathJax is desirable. Sławomir Biały (talk) 14:48, 20 November 2010 (UTC)

I think the most important thing is to offer a unified approach on the editor side. Meaning editors should be able to the <math>-tags and Latex notation inside for all their formulas. The current mix of between latex and html to make it look slightly better is imho a big mess and unnecessarily complicated to editors, who need to able to deal with latex and html instead latex only.--Kmhkmh (talk) 20:40, 20 November 2010 (UTC)

ƒ should be italicized in order to be consistent with TeX style. Michael Hardy (talk) 03:15, 21 November 2010 (UTC)

Rendering equations as black text on white, or not as it happens

I just want to show an example that demonstrates the fact that actually Wikipedia math images are not rendered as black text on a white background. By taking the equations from a thread on the laogarithm talk page and placing them on a colored background you can see that some of the images are black text on white, but others are black text on a transparent layer. I don't know why Wikipedia mixes and matches like this although I guess that all images used to be black-on-white, but at some point the developers decided to use transparent images instead, but images that have already been rendered in the past are not re-rendered but cached versions are used.

|z_1 + z_2| = 5

but this

|z_1 + z_2| = 5 \,

which is indeed the same as this

\displaystyle |z_1 + z_2|=5

It's why after a few years of experimentation black text on a white background is used everywhere, and why we have a guideline, WP:COLOR.

(blogb(x))p = bp · logb(x).
(b^{\log_b(x)})^p = b^{p \cdot \log_b(x)}
Example 1:
The following formula relates the logarithm of a fixed number x to one base in terms of the one to another base:
 \log_b{x} = \frac{\log_k{x}}{\log_k{b}}.\,
This is actually a consequence of the previous rule, as the following proof shows: taking the base-k-logarithm of the above-mentioned identity,
x = blogb(x),
logk(x) = logk(blogb(x)).
Example 2:
Calculation of powers are reduced to multiplications and look-ups by
cd = b(logbc) · d.
Divisions and roots are also covered by these two techniques since
\frac c d = c \cdot d^{-1} = b^{\log_b c - \log_b d}
\sqrt[d]{c} = c^{\frac 1 d}.
Furthermore, the math-equations are much easier to create, edit and maintain. Going back to your example, I think that maintaining
(''b''<sup>log<sub>''b''</sub>(''x'')</sup>)<sup>''p''</sup> = ''b''<sup>''p'' · log<sub>''b''</sub>(''x'')</sup>
producing (blogb(x))p = bp · logb(x),
is much harder and tiresome than
<math>(b^{\log_b(x)})^p = b^{p \cdot \log_b(x)}</math>
producing (b^{\log_b(x)})^p = b^{p \cdot \log_b(x)}.

Also from this f discussion: Take a look: f,\ g,\ h (compare that to ƒ, g, h) (talk) 23:31, 20 November 2010 (UTC)

As commented on this topic, since rev:59550 we can do the following when dealing with backgrounds:

Background Wikicode Rendering (in PNG)
White e^{i \pi} + 1 = 0 e^{i \pi} + 1 = 0\,\!
\definecolor{orange}{RGB}{255,165,0}\pagecolor{orange}e^{i \pi} + 1 = 0 \definecolor{orange}{RGB}{255,165,0}\pagecolor{orange}e^{i \pi} + 1 = 0\,\!
Orange e^{i \pi} + 1 = 0 e^{i \pi} + 1 = 0\,\!
\definecolor{orange}{RGB}{255,165,0}\pagecolor{orange}e^{i \pi} + 1 = 0 \definecolor{orange}{RGB}{255,165,0}\pagecolor{orange}e^{i \pi} + 1 = 0\,\! (talk) 02:16, 21 November 2010 (UTC)


I Agree with the use of mathjax on Wikipedia. This was recently also requested at this topic on pt.wikipedia.

See also other topics about this:

Helder 15:11, 20 November 2010 (UTC) —Preceding unsigned comment added by (talkcontribs)

  • Support. MathJax may be imperfect, but it is far better than the existing options.—David Eppstein (talk) 15:40, 20 November 2010 (UTC)
  • Support. Anything's better than the current system.--Dark Charles 19:46, 20 November 2010 (UTC)
  • Support. I have only passing familiarity with MathAjax, but as the others have said, it looks like a tremendous improvement over what we have. JeffConrad (talk) 22:11, 20 November 2010 (UTC)
Please have a look at my MathJax extension. Thanks for your interest, Nageh (talk) 20:08, 20 November 2010 (UTC)
PS: Instructions: (1) Install the TeX fonts from the MathJax package into your system folder. (2) Add the line "importScript('User:Nageh/mathJax.js');" to your Special:MyPage/vector.js (or whatever skin you are using) file. (3) Set the Appearance to "Leave it as TeX" in your Preferences. Done. Sorry for making you install the fonts. Nageh (talk) 20:18, 20 November 2010 (UTC)
Well, we can't ask users in general to download fonts. For this to be a general solution, it has to "just work" in modern browsers with no intervention on the user's part. I assume this will require developer intervention; is your plugin a step towards what they would have to do? --Trovatore (talk) 21:17, 20 November 2010 (UTC)
Yes, a big part of what makes MathJax good is that it "just works" with no user intervention required. Articles that can only be read by following special instructions would be a step even farther backwards. —David Eppstein (talk) 21:33, 20 November 2010 (UTC)
I am fully aware of the caveat. I was not suggesting this as a general solution but was offering it for your convenience, to see how it works. Btw, the immediate response from other users was that MathJax is still slow, so maybe you still want to wait a bit before asking for an official extension. Or may be not. It's all your decision. (FYI, there is no intrinsic reason why I currently require you to to download fonts. It's just that I find it tedious uploading each single font file manually. Just give me some time and it will work without user invention.) Thanks for the feedback anyway. Nageh (talk) 22:34, 20 November 2010 (UTC)
On second thought, I actually cannot upload fonts because they are none of the permitted file types. Google is providing a hosting service for web fonts, so it could work that way. On the other hand, there is nothing wrong with asking for official support of MathJax on Wikipedia. Nageh (talk) 10:42, 21 November 2010 (UTC)
  • (edit conflict) I'm not sure what we're voting on but my experience of MathJax as it works now is it's too slow. I had it enabled for a while and though it worked well enough except for a couple of oddities it was very slow: the page would load and then it would take a few seconds to run the script, announcing that it was doing so even on pages with no formulae. It was far slower and more distracting than PNGs. This may be a function of browser, connection speed and page complexity, but I think it needs to work on WP like the current renderer, not on users' machines so the SVGs are generated on the server, served like PNGs so they load as quickly as them, can be cached etc. --JohnBlackburnewordsdeeds 22:38, 20 November 2010 (UTC)
Comment: I think as a starting point it would help to have it as an option so people can opt in to test the results. A possible problem is that, while I have no idea what technical issues might arise with installing it on WP servers, I imagine there must be some and that means man-hours dedicated to getting it working. Unlike the other sites mentioned above, WP is not primarily a math site and keeping mathies happy is not the highest priority. But I agree that current system, while it may have been a good solution 5 years ago, is becoming obsolete as many browsers can now support MathML etc. It's probably going to take some dialog between the developers at WikiMedia, who will know what is possible and feasible, and people like us, who know what the problems are, before there can be some movement in the right direction.--RDBury (talk) 23:05, 20 November 2010 (UTC)
From my point of view rendering TeX as MathML would be preferable to images given the rendering problems described in the above section "Rendering equations as black text on white, or not as it happens" (talk) 23:22, 20 November 2010 (UTC)
As I commented here I don't see this as a problem as WP for good reasons of readability uses black text on a white background. If we had varying colour backgrounds I'm sure the problems visible above (in all the PNGs) would be fixed. As we have white backgrounds it's not an issue and has probably gone unnoticed by most editors because of that.--JohnBlackburnewordsdeeds 23:33, 20 November 2010 (UTC)
Yes, likely gone unnoticed by most. I can't help but notice because the transparent images don't work at all on my computer unless I change the operating system settings every time I want to see what an equation says. (talk) 23:40, 20 November 2010 (UTC)

I support any process to improve the rendering of mathematics on wikipedia and it looks like mathjax is the current strongest contender although the slow speed is concerning and could be a show stopper. Actually getting it enabled on en wikipedia will be a long progress and we would need to present a good argument to developers. Early steps would to get a test implementation on a foundation wiki possibly Wikimedia Labs and submitting a bug on Also compiling a list with defects in the current system would be a useful resource.--Salix (talk): 21:06, 21 November 2010 (UTC)

I'm in. But let me lay out what I think would be the best approach. As many have pointed out, MathJax is currently rather slow on maths-heavy pages. The devs are currently investigating several options for improving the current slow typesetting process, and there is some good indication that progress can be made. I would not suggest approaching Wiki developers prematurely, or the baby may be dead in the water. There is the possibility to test MathJax using the user extension that I provide, at the cost of installing 22 fonts into your system folder. (Less than I thought.) You may then decide when it is ripe to be rolled out officially. Or we may test an implementation right away on Wiki Labs. I don't know which of these two options is the better approach. Does it require begging the MediaWiki devs and them maintaining the extension? That could turn out as a showstopper. If a user space extension is equally accepted then all we'd need from the devs was to install the fonts on the Wikipedia/MediaWiki servers. Do you think this could be done? Anyway, both approaches could be suitable for a test phase before MathJax support is rolled out with an official extension. Nageh (talk) 22:53, 21 November 2010 (UTC)
Well, I think the endgame needs to be complete replacement of the PNGs, not as an option but as the default. Otherwise we can't get rid of the inline HTML. But I'm fine with doing it in stages if that will be easier to sell.
Someone mentioned the idea of having the servers generate the MathJax SVGs and cache them, as is now done with the PNGs. Is that technically feasible? --Trovatore (talk) 23:00, 21 November 2010 (UTC)
MathJax currently support four kinds of fonts: TrueType (.oft), which should be displayable on any PC, Embedded OpenType (.eot), which is a compressed format and only works in IE, SVG, which is the only method that currently works on iPad/iPhone, and at last, an extensive set of PNG characters. We don't want PNG, so all we need is 3 times 22 font files to be stored on a server. All that is required is a little help of some Wiki dev to upload them on Wiki web space. Nageh (talk) 09:00, 22 November 2010 (UTC)
I meant the SVG for a whole formula, not for a single character. --Trovatore (talk) 17:55, 22 November 2010 (UTC)
I was not suggesting caching single characters either. My point was that MathJax doesn't deliver a single SVG per formula, so you'd have to render them on an output canvas, but since that isn't Postscript like for TeX I figure this would be mean quite some extra work to implement.
And I don't like the solution anyway, because it's just half-way between what we have now and what would be a good solution. Nageh (talk) 18:31, 22 November 2010 (UTC)
OK, I'm interested; what would be the good solution? --Trovatore (talk) 18:50, 22 November 2010 (UTC)
A good solution would be one that actually uses a font(!) as opposed to some graphics. Because you can't copy graphics as text(!), and because you encounter the same problem as now where you don't know the correct baseline of the rendered text. (Well, it would be technically possible to solve the latter, but only with considerable extra complexity.) Nageh (talk) 19:25, 22 November 2010 (UTC)
Hmm, but are we talking about making the client machine run TeX implemented in javascript? That could be a "good" solution if the client were fast enough that you didn't notice it. Unfortunately current improvements in silicon are not really being directed to faster clock speeds on the desktop, certainly not at anything like historical rates.
Or do you mean an adhoc "font" that makes a whole formula into a single "character"? --Trovatore (talk) 20:07, 22 November 2010 (UTC)
This is exactly what MathJax is doing, typesetting TeX in real-time. And yes, it's a bit slow right now, but it is not because of limitations in current silicon technology but mostly because of browser page rendering issues (that (hopefully) can be worked out). This is all possible because only a limited subset of TeX is supported, namely what is required to set maths formulas. HTH, Nageh (talk) 20:33, 22 November 2010 (UTC)
It might be possible to get the Latex to MathML conversion to happen during the preprocessing and the resulting MathML (or HTML-CSS equivalent) cached. This would require server-side JavaScript. I'm not quite sure when the speed problems are occurring its either in the conversion or actually rendering the HTML-CSS version, for a simple equation like sqrt(x) it uses 18 span's to get the spacing right, so for a complex page its going to be a lot of elements to render.--Salix (talk): 09:30, 23 November 2010 (UTC)
Offtopic: I though you would find this to be interesting: Sparklines on Protovis. The tool inserts the images inline, in a way similar to what Trovatore was talking about (I think). Helder 13:24, 23 November 2010 (UTC)
  • Comment. One of the good things about Wikipedia's system is that it "just works" out of the box. There is no need to download special fonts, or to wait five minutes for your browser to get whatever it needs to display a page properly. Also, although the displayed math equations are not pretty, they display with nearly 100% reliability. My experience with MathJax has been that when it works, it looks very good, but some implementations are broken or unreliable. For instance, on my current browser, Terry Tao's blog looks fine, but the equations at mathoverflow don't display at all. If there were reasonable assurances that MathJax would work with nearly the same level of reliability that we have now, I would support this change in a heartbeat. Right now, I would support testing a change to see if it is workable. Sławomir Biały (talk) 13:36, 22 November 2010 (UTC)

Back to the original discussion

The original question was: f or ƒ? We've gotten completely sidetracked, but I think the original question is still important. I think ƒ is inappropriate. It is displayed in a serif font, so it clashes with all the other text on the page. There are no corresponding special characters for g, h, and so on, so it will clash with all the other functions on the page. Finally, because ƒ is a Unicode character it may cause accessibility issues for users with older software. Just because it's there doesn't mean we should use it. Ozob (talk) 11:55, 22 November 2010 (UTC)

I prefer f. Jakob.scholbach (talk) 13:31, 22 November 2010 (UTC)
I prefer ƒ. I think this should be like most other such stylistic choices. The MoS has no mandate on one or the other. Also, is there any evidence that anyone has trouble seeing this symbol? I can see it on every browser that I have tried it with (probably the first time that has happened). Sławomir Biały (talk) 13:39, 22 November 2010 (UTC)
I don't have a preference and would go with whatever the underlying engine for the math-tags renders. If no math-tag is used Jakobs preference is definitely more editor friendly, i.e. many editors who don't want to bother which html oder sign coding will use that version anyway.--Kmhkmh (talk) 14:20, 22 November 2010 (UTC)
I strongly prefer f. Writing ƒ instead is a misguided attempt to fake a serif character in sans-serif text by using an unrelated character with a hook. It's wrong both typographically and semantically. I'd also like to remind that whatever is the outcome of this discussion, there is nothing in the concept of "U+0192 LATIN SMALL LETTER F WITH HOOK" that would make it automatically italic, just like Ƒ or Ɠ are not italic either. For ƒ, this italic appearance is only a legacy feature of some fonts, but other fonts (such as the one I am seeing here right now) display ƒ upright. Using an upright hooked ƒ instead of an italic serif character looks patently horrible. Thus, if this misguided replacement is to be used at all, it must be explicitly italicized as ƒ.—Emil J. 17:48, 22 November 2010 (UTC)
I think ƒ is self-evidently a workaround. As long as it's kept in mind that it's a workaround, and that it will go away once we have a global solution, I don't care that much whether it's used on an ad-hoc basis. But I wouldn't use it myself, for Emil's reasons. --Trovatore (talk) 17:58, 22 November 2010 (UTC)
I think we should go with f. As I alluded to above, f looks more like f than a looks like a, yet we presumably don't want to say editors should switch fonts to obtain a different a (and similarly for v and v). I think it would waste editors' time to make such changes in their editing habits. RobHar (talk) 18:57, 22 November 2010 (UTC)
I believe you mean to compare f with f\,\! and a with a\,\!. Indeed, a versus a\,\! is something we should be concerned with, rather than just assuming that the issue of "f" is insignificant because the issue of "a" is. Sławomir Biały (talk) 02:46, 23 November 2010 (UTC)
Yeah, thanks, that is what I meant; I think I currently have the setting "always show PNG" set, so things were showing up fine for me. I guess, I see these problems as things it would be nice to solve, but not worth a using a hack every time. I certainly don't want to have to avoid f and a. RobHar (talk) 05:41, 23 November 2010 (UTC)
Again, I strongly second Emil's comments. Diddling content for appearance is not only illogical and bad typographical practice, it's often shortsighted as well; if a user were to set his CSS to use a serif typeface (or for some reason Wikipedia were to make this the default in the future), the workaround might prove to be anything but. As I mentioned earlier, I would not hold much hope for an improved TeX rendering, be it MathAjax or something else, to solve the problem. Until Wikipedia uses the same typeface for running text and mathematical material (like almost all other publications), there will be some obvious typographical mismatches. I think we minimize those mismatches by at least using the proper characters. JeffConrad (talk) 23:09, 22 November 2010 (UTC)
If anything, this is an argument that the characters used to typeset mathematics should not be dependent on user font settings (see related thread below). A special unicode character that typesets in nearly the same way as our latex->PNG conversion engine seems to settle this problem. Sławomir Biały (talk) 02:42, 23 November 2010 (UTC)
I have difficulty following this logic; why should a given character always look the same? Seems to me that a character should be represented as is appropriate for a given font, which is a collection of glyphs designed to be used together. If it is important that mathematical material be independent of user settings, this can be accomplished using <math> ... </math> tags. Of course then, the mathematical symbols may clash with the running typeface, as we've previously discussed. The uppercase I, in the discussion below is admittedly a special case; in most sans-serif typefaces, it's tough to distinguish from a lowercase l. There is no such problem with lowercase f; though it may not look the same in serif and sans-serif typefaces, there's no problem recognizing it. JeffConrad (talk) 04:23, 23 November 2010 (UTC)
The problem with the PNGs is not that they're in a different typeface from the prose. Why should math use the same typeface as prose? The problem is that they're (i) just plain ugly, even in and of themselves, and (ii) not sized correctly for inclusion in running text. --Trovatore (talk) 04:43, 23 November 2010 (UTC)
On top of which they don't work well for background color changes (as the above discussion indicates), they don't automatically resize when one resizes the browser's text, and they don't allow text from within a formula to be copied and pasted. All of which is better in MathJax. Formatting math manually in html also fixes most of these issues, but maybe not your first point about being ugly. —David Eppstein (talk) 05:06, 23 November 2010 (UTC)

Staying with the original issue, why should math use a different typeface from the running text? If it does, either
  1. The symbols in running text will differ from those in displayed material, or,
  2. If the symbols in running text match those in displayed material, the symbols will clash with the typeface in the running text.
In the past, using a separate mathematics font is not without precedent; early implementations of AT&T's troff did so. But in that case, the main reason for the special font was the inclusion of mathematical symbols not present in the regular fonts. Most current fonts are not limited to 256 glyphs, and include the vast majority of mathematical symbols, so the need for a special mathematical font is considerably reduced. I'll return again to Bkell's reference to professionally typeset mathematical publications, in which the typeface for mathematics usually matches that in the running text. Again, if running text and mathematics use different typefaces, there is bound to be a mismatch somewhere. If we choose characters rather than glyphs, we at least have some degree of self consistency within a given class of material. JeffConrad (talk) 07:53, 23 November 2010 (UTC)
But, in reality, Wikipedia already does use a different symbol\, \text{set}\, to typeset mathematics! Sławomir Biały (talk) 20:51, 28 November 2010 (UTC)

I'd like to ask again, are there any documented accessibility issues of using ƒ as opposed to f? Or is this just wishful thinking for the sake of trying to win an argument? Sławomir Biały (talk) 20:53, 28 November 2010 (UTC)

No, they are at the moment all hypothetical. Someone from WP:ACCESSIBILITY WP:WPACCESS could probably give us a clearer answer, so I've posted a question there. Ozob (talk) 21:13, 28 November 2010 (UTC)

Update of MoS

In accordance with this discussion and the related discussion at WT:WPSTAT, I have updated the math MoS to explicitly disallow use of ƒ for a function. Ozob (talk) 14:43, 28 November 2010 (UTC)

I don't really believe there is consensus for this change. Several editors have stated a preference of ƒ over f. Most of the dissenting voices have failed to argue a coherent position (e.g., nonexistent accessibility issues brought up, etc.) At any rate, the change to the MoS seems premature to me, particularly given that many articles do use ƒ, and have for quite a long time. Sławomir Biały (talk) 20:45, 28 November 2010 (UTC)
As a user of the screen reader JAWS, I'd prefer the use of the character "f". The "ƒ" symbol reads as floren, but it's only read in the text when the punctuation mode is set to all (i.e. it's treated the same way as sighted people would read a comma in a block of text except without the associated pause). This setting can be changed, but not many people would know how to do it. Graham87 02:27, 29 November 2010 (UTC)
User:Dodoïste indicated on WT:WPACCESS that NonVisual Desktop Access reads the character as "letter 1 9 2", so it does not work there, either. My conclusion from this is that ƒ is not accessible for disabled users. Also, my reading of the above discussion is that there editorship has a slight preference for f over ƒ: A few more editors expressed preference for f than ƒ, but not many, and there were several who were neutral or indifferent. Putting these two together, I think that the consensus is to forbid ƒ. I'm not going to update the MoS on this basis yet, though, since I seem to have misread the situation last time; I'd prefer to have other opinions first. Ozob (talk) 03:11, 30 November 2010 (UTC)
I am generally opposed to mandatory solutions like the one proposed, particularly if they are likely to bring many existing articles suddenly into non-compliance. But if it comes down to a legitimate access issue, then I withdraw my objection to the change. Sławomir Biały (talk) 13:48, 30 November 2010 (UTC)
But how are LaTeX formulas read using NonVisual Desktop Access? I am very much against this change. Both f and ƒ are acceptable, although my preference is for the latter. Choices should not be forced onto people. If there's a problem with some programs not reading it correctly then it's those programs that need changing and not the millions of articles on Wikipedia. Fly by Night (talk) 16:01, 30 November 2010 (UTC)
My guess is that the NonVisual Desktop Access will read the alt text of the typeset formula, which is the verbatim LaTeX code.TimothyRias (talk) 16:09, 30 November 2010 (UTC)
So LaTeX would be total nonsense to someone using NonVisual Desktop Access. Should we stop using LaTeX too? If not, and they can understand verbatim code then letter 1 9 2 shouldn't be a problem. Fly by Night (talk) 17:33, 30 November 2010 (UTC)
I don't know whether you actually mean such an utter nonsense seriously. No one's going to understand "letter 1 9 2" as it's a meaningless code, whereas mathematical audience can be safely assumed to understand LaTeX.—Emil J. 17:51, 30 November 2010 (UTC)
I do mean every last word of it, and please don't be so bad tempered and hostile. If someone can understand LaTeX code read to them, verbatim, and they can understand what that means, then hearing letter 1 9 2 will not be a problem at all. If you know what
backslash int underscore zero caret one backslash sine open bracket ex close bracket de ex
means then letter 1 9 2 would be a piece of cake. Are you suggesting that someone using NonVisual Desktop Access could learn LaTeX so well without being able to understand what letter 1 9 2 means after being told once or twice? Fly by Night (talk) 18:13, 30 November 2010 (UTC)
I can imagine that in isolation, someone could learn what "letter 1 9 2" means, even if it is completely unmotivated. However, that's not how it works. This is simply a default rendering of Unicode characters the system does not have a special rule for, and there is no way of recognizing "letter 1 9 2" among dozens of symbols like "letter 1 0 A C", "letter D 1 2", "letter 2 1 0 9", and so on (I'm making the numbers up, I didn't look up what these characters really are). There are some 100,000 of code points, and you can't expect anyone to learn them. Moreover, we should not force anyone to learn such a code just to be able to read Wikipedia. In contrast, LaTeX is a widely used standard for representing mathematics, and a mathematically inclined reader will be familiar with it at least to some extent, they do not have to learn it because of Wikipedia.—Emil J. 18:36, 30 November 2010 (UTC)
I agree with what you're saying in isolation, and you make a good point. However, in WP:NOT PAPERS it clearly says that "A Wikipedia article should not be presented on the assumption that the reader is well versed in the topic's field". Assuming that everyone reading a maths article is well versed in LaTeX goes against this policy. I didn't meet LaTeX as an undergraduate. I would guess that 95% of undergraduates can't write in LaTeX, nor understand LaTeX code. So we're left with the postgraduates, that make up a tiny fraction of the mathematically educated world. In short: if we force people to use f then we need to stop people using LaTeX. Fly by Night (talk) 20:55, 30 November 2010 (UTC)
While "letter 1 9 2" is an unhelpful rendering, JAWS' reading it as "florin" is perfectly correct, as the letter is the florin sign, despite being abused for f in some Wikipedia articles (no, not millions, it's hundreds at most). If we want it to be read as an f, we should write it as an f.—Emil J. 16:21, 30 November 2010 (UTC)
To reiterate: it's not an abuse. That claim has already been refuted. This is one of the stated intended uses of the symbol, per [4] Sławomir Biały (talk) 16:46, 30 November 2010 (UTC)
No, this is just the result of total confusion on the part of some standardization body. As a mathematician, you should be aware that there is no such thing as a dedicated "function symbol". Functions are denoted the same way as any other mathematical objects, i.e., by Latin letters, Greek letters, and a wide selection of odd symbols, potentially with subscripts, diacritics, and other decoration. If you'd want to interpret the comment you linked to on its face value, you'd have to write ƒ = ƒƒ instead of f = gh, ƒ(2x) = 2ƒ(x)ƒ(x) instead of sin(2x) = 2sin(x)cos(x), and so on. It simply does not make any sense.—Emil J. 17:13, 30 November 2010 (UTC)
To clarify, the confused standardization body is likely not the Unicode Consortium. These comments after = you see in the pdf are informative alternative designations generally taken from older legacy encodings for compatibility purposes. The Unicode designation of the character is only the thing in all caps.—Emil J. 17:29, 30 November 2010 (UTC)
This is a completely specious argument. You know perfectly well that neither I nor the Unicode consortium would advocate using just one symbol to denote a function. However, that one of the stated purposes of the symbol ƒ is to denote a function completely obviates your point that using it in this manner constitutes an "abuse". Sławomir Biały (talk) 17:26, 30 November 2010 (UTC)
I'm with Sławomir on this one. On another point: I was under the impression that ƒ was prefered over f in the WP:MOSMATH. I've seen editors change f to ƒ and cite WP:MOSMATH. Fly by Night (talk) 17:33, 30 November 2010 (UTC)
The upright ƒ is not an option and should not be used under any circumstances, even if your font happens to display it as if it were italic. The discussion is whether to allow italic ƒ in place of italic f.—Emil J. 17:48, 30 November 2010 (UTC)
I didn't say it was; did I?! Less of the smoke and mirrors, please! I just forgot to put it into italics, that's all. Fly by Night (talk) 18:07, 30 November 2010 (UTC)
@SB: no, I don't know any such thing, that's only your interpretation. My interpretation is that it was invented by a misguided non-mathematician who was not even aware that there can be other functions than f, and who confused a font difference with a special symbol.—Emil J. 17:56, 30 November 2010 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────To Fly by Night: When I started editing Wikipedia, I knew nothing about Tex (although I was familiar with computer programming and mathematical notation). However, I picked it up rather quickly, especially after I became aware of Help:Displaying a formula. So I do not think that lack of experience with it is a major stumbling block, except perhaps for people with disabilities such as the blind person mentioned above. JRSpriggs (talk) 00:29, 1 December 2010 (UTC)

That's exactly the case I had in mind. People that are using programs that read the page out to them. They hear letter 192 or florin, and strange things like that, when they come across ƒ. People were saying we should stop using ƒ for that reason per WP:WPACCESS. I argue that if we stop using ƒ then we should stop using LaTeX. Of course I don't think we should get rid of LaTeX code, and we shouldn't stop using ƒ. Fly by Night (talk) 13:26, 1 December 2010 (UTC)
[5] says that JAWS will read alt text. So anything between <math> tags will be read aloud as raw LaTeX. I don't know exactly how JAWS works, but I'm guessing from Graham87's comment about not reading punctuation by default that something like \textstyle\int_0^1 f(x)\,dx (entered as \textstyle\int_0^1 f(x)\,dx) would likely be read as "textstyle int zero one eff ecks dee ecks". I think that's pretty intelligible. I'm not sure whether it's better or worse if the underscore and the caret are read; it's certainly worse if the backslashes and the comma are read. More complicated expressions may fail to be intelligible: If the underscore and caret are not read, then g_{ik}g^{kj} = \delta_i^j would become "gee eye kay gee kay jay equals delta eye jay", and that loses too much context to make it possible to reconstruct the original formula (since it could just as well have been g^{ik}g_{kj}). In this example, it may be possible to work around this by writing the whole formula in HTML (it depends on whether or not JAWS parses <sub> and <sup> tags). Of course, the moment one tries to do anything more complicated then HTML is no longer an option.
I don't really see a good way around this. A screen reader needs context to understand what the written word is trying to communicate or else it may not be able to read correctly. If the screen reader speaks LaTeX and if it recognizes that we're writing LaTeX, then it'll do fine. If we code everything in MathML and the screen reader speaks MathML, then it'll do fine. But either of these are very hard! And while reading the LaTeX aloud character by character is suboptimal, I can't think of another easy solution.
Finally, the fact that LaTeX does not work well with screen readers is irrelevant to the discussion about ƒ. With ƒ, we have an alternative: Plain f. The only cost to mandating a plain f is the work of converting pages that currently use ƒ. Whereas the cost of a solution that reads arbitrarily complicated formulas is prohibitive. Ozob (talk) 03:20, 1 December 2010 (UTC)
It's not irrelevant at all. If we should get rid of ƒ because people don't understand what letter 192 means, or what florin means then we should get rid of LaTeX for the same reason. However, as I said above, we shouldn't stop using LaTeX and we shouldn't stop using ƒ. Fly by Night (talk) 13:29, 1 December 2010 (UTC)
The reason you shouldn't use ƒ is because nobody in their right mind will understand that "florin" or "letter 192" means the same thing as "f", because, well, they do not. The solution to TeX display using a different font, than the default stylesheet of wikipedia, is not hacking around with unicode characters that look like certain characters. (For one this undermines future compatibility, if the math font used is changed or becomes customizable.) If you really want to name a function ƒ, then your TeX equations should also use the florin symbol (that doesn't seem to be supported by the TeX engine run on the wikipedia servers.)
As for screen readers read alt text. Having TeX code read out loud is probably the best bet of a blind person understanding a computer read math equation. It is probably easier to understand (and more universal) than having the various HTML hacks read out loud.TimothyRias (talk) 13:48, 1 December 2010 (UTC)
Did you read the thread? It really comes across like you haven't read anything that's been written. Myself and others have already addressed the points you make. For example, florin is the name of the symbol ƒ, and LaTeX isn't standard. Most people that read Wikipedia maths articles are non-editors without postgraduate eductation. Meaning they won't have met LaTeX and they won't have opportunity to pick it up. Assuming that LaTeX is standard goes against WP:NOT PAPERS where is says, and to repeat, "A Wikipedia article should not be presented on the assumption that the reader is well versed in the topic's field." Fly by Night (talk) 15:27, 1 December 2010 (UTC)
Yes, I read the thread. Your suggestion is a huge violation of WP:AGF and WP:CIVIL, for which I would appreciate an apology. Yes, "florin" is the name of the ƒ symbol, that is because it is the florin currency mark. Depending on the font that is used to render the wikipedia page, this will resemble the the italic f of the computer modern serif font, or not. (For example, in the font that I get in the edit window, it doesn't look like the computer modern serif italic f at all.) The point is that this symbol is not an f, while the symbol presented by the LaTeX is an f.
This can lead to future compatibility issues, when for example wikipedia would introduce the option for users to give a preference in which font LaTeX formulas are rendered (this is not unlikely to happen if something like mathajax is implemented). Some people might for example choose to use a sans-serif font to match the default on screen font, this will make articles using the "florin" hack look very silly. Thus, for sake of future compatibility alone it is better to just stick with using something that is an "f" to represent an "f". Using a different symbol, that might look slight better on your screen, or even in general for current browsers, just is a bad idea for building article that are future proof.
As for LaTeX being standard or not. Of all the ways of romanizing mathematics it is likely to be the most standard. That is, there is probably no other way of representing an complex equation by a string of Roman characters that is more likely to be understood by a blind person listening to it, except maybe full transcription of the formula. It certainly is more likely to be understood than the series of HTML hacks used to represent equations in pure HTML.TimothyRias (talk) 17:13, 1 December 2010 (UTC)
Yeah, I agree. I don't want LaTeX to be done away with. I was saying argument for getting rid of ƒ, per WP:WPACCESS, applies directly to LaTeX and so if we get rid of ƒ then we should get rid of LaTeX. We wouldn't ever get rid of LaTeX, so we shouldn't get rid of ƒ. However, the compatibility issue that you mention is a very good point indeed, and I'm inclided to agree. As for your apology: please, less of the drama. You can't throw WP:CIVIL and WP:AGF at me just because I ask you if you've read something. Fly by Night (talk) 17:35, 1 December 2010 (UTC)
It is not correct to compare ƒ to LaTeX. We can replace ƒ with f; it works for everybody and requires no software changes. This is why ƒ is different from LaTeX: What do you propose to replace LaTeX with? Ozob (talk) 03:41, 2 December 2010 (UTC)
I've stated three or four times that I don't think we should get rid of LaTeX. Let me repeat my very last post for you: "I don't want LaTeX to be done away with. I was saying [that the] argument for getting rid of ƒ, per WP:WPACCESS, applies directly to LaTeX and so if we get rid of ƒ then we should get rid of LaTeX. We wouldn't ever get rid of LaTeX, so we shouldn't get rid of ƒ." P.S. I won't be taking any more part in this thread. It's going round in circles. I guess it must have got too long. But people seem to be making the same points over and over; points that have already been address. I'm getting bored writing the same things again and again. Let me know what the outcome is over on my talk page. Cheers. Fly by Night (talk) 12:30, 2 December 2010 (UTC)
And as severe people have pointed out to you, that argument does not fly, because for LaTeX there is no alternative, whereas there is a perfectly good alternative to using ƒ, namely using f.TimothyRias (talk) 13:05, 2 December 2010 (UTC)

{{math}} template

Formulae are traditionally typeset if a serif font (just as Latex does). Even if the rest of the print is in a different typeface like sans-serif, formulae still use serif, for the exact same reason this whole discussion point out: to properly display the various symbols that are distinguishable (where sans-serif has major problems). This is also the exact reason {{math}} exists. It defaults to serif, prevents line wrapping and is finetuned to fit around the text in every skin. If you all were to use {{math}} consistently, the entire discussion above would be solved. f, g, h becomes f, g, h, and (blogb(x))p = bp · logb(x) becomes (blogb(x))p = bp · logb(x).

Now, some may not like the mix of various typefaces, but I see no less of that in my math books, so this argument holds no water. This is a case where functionality must have priority over aesthetics. In order to properly display formulae, the simplest solution is often the best one. There is an ongoing discussion regarding inline formulae on Wikipedia talk:Manual of Style (mathematics)#Request for comments: serif vs. sans-serif. EdokterTalk 14:32, 4 December 2010 (UTC)

Request for comment concerning formatting of standalone equations

It seems November is the month of formatting questions. After an elongated discussion with DVdm at Talk:Logarithm, which did not come to any consensus, we agreed to put this request for comment. The questions is this (I should probably disclose that DVdm is in favor of the following suggestion, I'm not):

Should WP:MOSMATH be changed by adding to it (or merging) the text in the grey box below?

Mixing HTML and TeX markup for standalone mathematical formulas (as opposed to inline mathematical expressions) in the same article is not acceptable, regardless of the formatting of inline expressions (which should generally be formatted using HTML, if possible). For example, the following example has to be replaced by the one below.

Rational numbers depend on natural numbers. For any natural numbers p and q we write
x= \frac p q\,
for the fraction of p and q. In this case, the equation
x · q = p
holds, but
xq · p.


Rational numbers depend on natural numbers. For any natural numbers p and q we write
x= \frac p q\,
for the fraction of p and q. In this case, the equation
x \cdot q = p
holds, but
x \neq q \cdot p. \,

Thanks for any input. Jakob.scholbach (talk) 21:11, 23 November 2010 (UTC)

Would I be right in saying that the question is about whether some formulas can be displayed in HTML and others in LaTeX? My personal preference is that in-line expressions should use HTML and displayed expressions should use LaTeX. Fly by Night (talk) 21:25, 23 November 2010 (UTC)
Yes, this is exactly the question. Jakob.scholbach (talk) 10:03, 24 November 2010 (UTC)
Although I would not word it like that it needs clarifying. Right now both WP:MOS and MOS:MATH mention, as the very first item and in the first paragraph respectively, the need for consistency. I interpret that as meaning articles should use LaTeX or HTML for standalone formulae but not both. This often means LaTeX should be used as it can handle all formulae, while HTML can't, and on a long article like Logarithm there are usually formulae only possible with LaTeX. There are other arguments, such as readability of small subscripts, especially subscripts of superscripts, but the consistency one is the main one.--JohnBlackburnewordsdeeds 21:43, 23 November 2010 (UTC)
No, the text in the above box is too long and detailed. All that's needed is a single statement, perhaps with references to other policies like those above. What needs clarifying is this:
"If you find an article which indents lines with spaces in order to achieve some formula layout effect, you should convert the formula to LaTeX markup."
I read this as meaning any standalone formula, i.e. on its own line and indented with spacing characters (one or more ':' or '*') needs converting to LaTeX. Other editors interpret it as meaning formulae made into pre-formatted text by prepending a space should be made into LaTeX. But that such formatting should not be used is not particular to maths articles, though the main MOS seems silent on it perhaps as it's too obvious.--JohnBlackburnewordsdeeds 22:39, 23 November 2010 (UTC)
I agree with DVdm that the important point is covered by the consistency guideline and that standalone formulae in a given article should all be either HTML or LaTeX without a mixture. The Logarithm article seems to require LaTeX for all standalone formulae. However, since not all editors of math articles reach that conclusion from the consistency requirement, it probably should be made explicit.Support.and.Defend (talk) 03:42, 24 November 2010 (UTC)
  • Comment #2. I just realised that Jakob seems to have intended this to be an RFC (Request For Change) whereas on the Logarithm talkpage I always interpreted the acronym RFC to stand for Request For Comments, as can be seen both here and in the description of {{Rfctag}} with which this section is tagged. A request for MOS-change should probably be done on the MOS talk page. Anyway, if this is indeed the case (request for Change), and the idea was that only a change to the MOS would allow other editors to replace the HTML-equations in the Logarithm article with TeX, then I strongly distantiate myself from this. On the talk page I agreed with a request for comments about a point of discussion. I did not agree with filing a request for Change to the MOS. So, Jakob, if that was indeed your idea, then in my opinion that would be an extremely wp:POINTY and wp:OWNY way to go forward: like Jakob.scholbach's HTML-equations can be replaced with TeX if and only if the community is prepared to change the MOS to explicitly reflect a particular editor's interpretation. I'm sorry to put it this way, but I sincerely hope that this is not what you had in mind here. Can you please comment on this? Thank you. DVdm (talk) 06:28, 24 November 2010 (UTC)
Yes, DVdm and JohnBlackburne, we do have the guidelines you cite, but there was no consensus that these guidelines actually prescribe using TeX in all standalone equations, because the inline math expressions are formatted using HTML anyway, so the strive for 100% consistency is not satisfiable anyway. This is why we need to determine the community's consensus here.
@DVdm's 2nd comment: your accusation with a question mark is clearly unfounded. Nowhere do I talk about my (or anyone else's) equations. We had a long discussion, with you and John just digging your heels into the MOS-consistency thing, but IMO unwilling to actually consider any further arguments. The only way to move on was this RfC[omment] and we all agreed on this. I fail to see how this is POINTY, especially given that the proposal I suggested exactly matches your formatting predilections. If this RfC yields consensus that the passage I suggested above is already contained in or subsumed by the total sum of other guidelines, we of course don't need to add it to MOSMATH. However, given the energy that went into the previous discussion, it seems wise to me to clarify the MOSMATH guidelines as much as possible. Jakob.scholbach (talk) 10:03, 24 November 2010 (UTC)
Again, for the so-many-th time, the idea is not to prescribe "using TeX in all standalone equations", but to prescribe "using one type of rendering in all standalone equations, regardless the rendering used in inline equations, which can remain in HTML." I do not strive for that kind of 100% consistency.
Regarding the second comment, it was not meant as an "accusation with a question mark". It was a question. If the answer would have been positive, a formal accusation of point and ownership would have followed elsewhere. In other words, if the community is not prepared to change the content of the MOS like you suggest, I do not expect you to use that failure as a reason to keep the article in its current state with a mix of standalone math equations. Thanks for having answered the question. DVdm (talk) 10:49, 24 November 2010 (UTC)
In general, like User:Fly by Night, I use HTML for inline and TEX for inset expressions and that's what I've seen in nearly every math article I've looked at, though there are probably some exceptional cases so being overly prescriptive is probably not a good idea. The reason it's more or less standard is because it looks about as good as we can get with the software we're using and everyone is used to seeing it that way. But I suspect that most readers won't notice the difference or care much if they do.--RDBury (talk) 10:59, 24 November 2010 (UTC)

Can we conclude from the comments here, higher up on this page at Wikipedia_talk:WikiProject_Mathematics#Formatting issues, and at talk:Logarithm that (1) there is no objection against bringing standalone equation TeX-consistency to the article Logarithm, and that (2) it is not clear whether changes should be made to guidelines? DVdm (talk) 18:54, 29 November 2010 (UTC)

I think we can conclude that people need to use their common sense when writing articles. Make it look as nice, and well ordered as possible. I recommend not using LaTeX in-line because it renders out-of-line and at a different size to all of the other HTML. If you have to use in-line LaTeX (because the symbol doesn't exist in HTML) then use \scriptstyle. If you're going to display a formula on a line on its own then I would recommend using LaTeX. HTML looks too small and too insignificant to warrant a whole line to itself; besides LaTeX was designed for mathematical text and looks much better than HTML. There's no need for any change in policy. Fly by Night (talk) 22:02, 29 November 2010 (UTC)
Whatever you do, don't use scriptstyle for inline math outside superscripts and subscripts! This was last discussed here — Carl (CBM · talk) 22:10, 29 November 2010 (UTC)
You seem to misunderstand what I said. I meant that as a last resort, if the symbols aren't available in HTML, then \scriptstyle is better than ordinary sized LaTeX. All of the examples in your link could have been typeset in HTML and so \scriptstyle would never have been used. For example x + 2ni is fine in HTML and is better than x + 2^{n_i} that you seem to suggest in the link. Just look at how out-of-line it is! In fact, the topic you linked to seems to also talks about HTML against \scriptstyle. Of course, if HTML can be used in-line then it should be used. I say HTML should always be used in-line when possible. If you have to use LaTeX then \scriptstyle is better than normal LaTex. You would only ever need LaTeX for the odd symbol, say \scriptstyle \mathfrak{X} which looks a lot better than \mathfrak{X}. Look how big it is compared to the rest of the text. It just looks silly. \scriptstyle \mathfrak{X} looks much better; although it's still out of line. HTML would be the best way, but that symbol doesn't exist in everyone's fonts, so LaTeX is the only option. If it's a longer expression then a displayed formula could be used. Fly by Night (talk) 22:28, 29 November 2010 (UTC)
Indeed, I do not plan to touch any inline equation. I intend to TeXify the standalone equations for consistency within the article. DVdm (talk) 22:29, 29 November 2010 (UTC)
And you have my !vote. Fly by Night (talk) 22:49, 29 November 2010 (UTC)
I agree with point (2) that no change of MOSMATH or whatever is mandated by this RfC. I disagree though with conclusion (1). We have had three additional people commenting, Fly by Night advocating for common sense (although personally prefering TeX), Support.and.Defend voicing support for a policy change towards LaTeX-ifying, RDBury saying that no prescription should be made (also personally preferring TeX). To me, this reads like there is (still) no consensus that HTML in standalone equations is "unacceptable". Therefore, there is no consensus that such a change has to be made. Jakob.scholbach (talk) 22:17, 29 November 2010 (UTC)
I know you disagree with (1), and you seem to be the only one. The only objection I have seen put forward is that you have put a lot of work and thought in it. Noting that a wp:consensus does not mean unanimity (see Wikipedia:NOTUNANIMITY#Not_unanimity), I think we have a consensus that there is no valid objection against (1). DVdm (talk) 22:29, 29 November 2010 (UTC)
OK, if Fly by Night also voices support for TeX in standalones, we have a 2 in favor of prescribing TeX in this case and 1 editor who does not want to prescribe it. Even though I don't think this is a qualified consensus (just 3 additional opinions), I no longer object to the changes proposed by DVdm. I removed the RFC tag. Jakob.scholbach (talk) 20:01, 1 December 2010 (UTC)
Jakob.scholbach: My preference for LaTeX is context dependent. I would avoid in-line LaTeX as far as humanly possible. There is no consensus that displayed HTML is unacceptable. As long as the article is consistent. It just looks rubbish. Many mathematicians of a certain age view LaTeX as one of the greatest inventions of their lifetime. And they do so for a reason. Using HTML in displayed formulas, IMHO, is like going back in time 40 years to the old typewriter written books. Fly by Night (talk) 22:47, 29 November 2010 (UTC)

I believe this discussion, and the discussion above about f or f (at least somewhat) are covered under MOS policy already: Wikipedia:Manual_of_Style_(text_formatting)#Words_as_words WP:MATH should either interpret that policy as applying in these cases or adopt a policy consistent with this policy.Greg Bard (talk) 02:10, 30 November 2010 (UTC)

Per Jakob's most recent comment, I will start standalone-TeXifying the Logarithm article section by section, in no particular order, although lazy-first springs to mind. Later I might go to one of the guideline articles and propose some slight amendment in order to avoid discussions like the above. DVdm (talk) 13:52, 4 December 2010 (UTC)

Speedy deletion of "Everest (mathematical competition)"

At User:Michael Hardy/Everest (mathematical competition) I've put a copy of an article titled Everest (mathematical competition) written by User:Luka.W.B. that was speedily deleted by User:JzG. Should it have been deleted? Michael Hardy (talk) 14:48, 3 December 2010 (UTC)

Definitely not a {{db-a7}} candidate. It "credibly indicate[d] the importance or significance of the subject" on the first line. In needed referencing, but wasn't a speedy candidate. Fly by Night (talk) 15:11, 3 December 2010 (UTC)
From WP:CSD#A7: The criterion does not apply to any article that makes any credible claim of significance or importance even if the claim is not supported by a reliable source or does not qualify on Wikipedia's notability guidelines. I tend to think that "one of the most famous mathematical competition in Georgia" is a credible claim of significance or importance.—Emil J. 15:17, 3 December 2010 (UTC)

I've undeleted it; those who want it deleted should take it to AfD. Michael Hardy (talk) 15:20, 3 December 2010 (UTC)

The person who initially deleted it has taken it to AfD. See Wikipedia:Articles for deletion/Everest (mathematical competition). Michael Hardy (talk) 18:56, 3 December 2010 (UTC)
Just to be clear, you are talking about the nation Georgia in Eastern Europe, not the state of Georgia in the United States. JRSpriggs (talk) 06:22, 4 December 2010 (UTC)

Symmetric product

The article titled Symmetric product of an algebraic curve is about something that one can apply to topological spaces generally, and maybe to other things as well. Should we make the subject broader and change the title, or should we have a separate article on the more general topic?

In the mean time, I've changed the page titled symmetric product from a redirect to a section on the symmetric part of a tensor, to a disambiguation page. Michael Hardy (talk) 18:10, 4 December 2010 (UTC)

Cyrus Pitt Grosvenor

New article about the Baptist preacher Cyrus Pitt Grosvenor. It says he contributed to the problem of squaring the circle, so people might want to look at it. Maybe the article even qualifies to be tagged by this WikiProject, with "field" set to mathematician. Yaris678 (talk) 20:50, 5 December 2010 (UTC)

I wouldn't think so. He was an anti-slavery, baptist minister; so I wouldn't call him a (professional) mathematician. The problem of squaring the circle is quite elementary, and I'd say it comes under recreational mathematics. Fly by Night (talk) 23:40, 5 December 2010 (UTC)
Even before it was proved that pi is transcendental? That being the case, should we add a bit of context to the article on Cyrus Pitt Grosvenor? Yaris678 (talk) 00:29, 6 December 2010 (UTC)

The problem of squaring the circle is very very far from elementary! Michael Hardy (talk) 03:10, 6 December 2010 (UTC)

To get the right answer is very far from elementary, but I think what FbN meant is that a lot of folks worked on it at a recreational-math level. It's quite true, they did (probably still do). They never really got anywhere, of course. Whether that makes them "mathematicians" or not is an interesting question, but I probably wouldn't add the WPM banner to their articles' talk pages. --Trovatore (talk) 21:10, 8 December 2010 (UTC)

Inversion (discrete mathematics)

Deletion of Inversion (discrete mathematics) is proposed. Opine here. Michael Hardy (talk) 04:58, 6 December 2010 (UTC)

Replace with a redirect to permutation inversion. --Matt Westwood 06:19, 6 December 2010 (UTC)
The AfD would be a better place to express your opinion. —David Eppstein (talk) 06:29, 6 December 2010 (UTC)


An IP has repeatedly added a mention of theorem called "G-Taylor", an apparent generalization of Taylor's theorem, to Taylor's theorem (e.g. [6]). I removed it once as, at first glance, it seemed probably not significant enough (yet?) to be included in that article. Anyone else have an opinion on this? Paul August 18:11, 6 December 2010 (UTC)

The article is published in a vanity press, so it's not appropriate for Wikipedia, even if we ignore the (obvious) WP:COI issues. CRGreathouse (t | c) 18:48, 6 December 2010 (UTC)

CRGreathouse, removed the mention of "G-Taylor", another (or the same editor using a different IP) added back the mention, I removed it again (asking for a discussion on the talk page) and the original IP has restored it again (with no discussion). Do other editors wish to weigh in? Paul August 18:08, 7 December 2010 (UTC)

I've filed a RPP; asking for temporary, semi-protection citing an edit war. ([7]) Fly by Night (talk) 18:31, 7 December 2010 (UTC)

The page has been protected for a week. The protecting admin said that one week would give enough time for the problem to be resolved on the article's talk page. I suggest we carry on this discussion over on the article's talk page. That way, we'll have something in writing, attached to the article itself, in terms of consensus. Fly by Night (talk) 13:52, 8 December 2010 (UTC)

Fine. I will copy the above discussion to Talk:Taylor's theorem. Paul August 19:49, 8 December 2010 (UTC)

Infinite expressions

We have an editor (and it's not the one you're probably expecting) who insists on taking a formalistic/logical point of view in Talk:Continued fraction; the specific issue is whether it's permissible to use the phrase "infinite expression" without first precisely and formally defining exactly what such a thing is. Additional opinions there might be helpful. —David Eppstein (talk) 21:14, 7 December 2010 (UTC)

I think it is safe to conclude that the aforementioned editor is I. To clarify, it is not that I am proposing that "infinite expression" should be defined in the Continued fraction article. I am proposing that if we are going to use a term like "infinite expression" in the article then the term "infinite expression" should be a well-defined mathematical concept. We appear to have converged to an agreement that "infinite expression" is defined by Euler, among others, as objects that are infinite series, infinite products, or infinite continued fractions. We have not located a formal definition of "infinite expression" that is more general. —Quantling (talk | contribs) 17:34, 8 December 2010 (UTC)
You have just used "defined by" and "Euler" in the same sentence. This really should be a tripwire; Euler didn't work that way.
You are insisting that every turn of phrase in a mathematical article be defined with mathematical precision, which is impossible and undesirable. Impossible because some phrases must be undefined terms, used to explain the others; undesirable because even the close approach to it which is possible would take one and a half volumes to state 1+1 =2. Some things in Wikipedia articles are common English, not to be held to question: we are not a textbook. Septentrionalis PMAnderson 20:53, 8 December 2010 (UTC)
To follow your example, I am not trying to suggest that an article on 1+1=2 should be one and a half volumes. However, as I think you would agree, I want whatever is in the 1+1=2 article to be true. The subtlety is what to do with a sentence like "1+1=2 is an example of a set theory gymnastics"? The term "set theory gymnastics" sure looks like technical jargon, but, in actuality, it is not a well-defined mathematical concept. Should we keep that sentence or not? This example is an exaggeration of the actual question above, but perhaps it gives a feel for the nature of my objection. —Quantling (talk | contribs) 21:11, 8 December 2010 (UTC)
The part about taking 1 1/2 volumes to set up the definitions needed to show with full mathematical precision and rigor that 1+1=2 is not a made-up or exaggerated example, by the way; see Principia Mathematica. —David Eppstein (talk) 21:22, 8 December 2010 (UTC)
If Euler is a tripwire in this case, would you agree that this is additional evidence that "infinite expression" is on shaky grounds as a well-defined mathematical concept? Do we really want the Continued fraction article to begin with "In mathematics, an infinite continued fraction is an expression that …" if infinite continued fractions are not (finite) expressions and there is no well-defined "infinite expression"? I want to find a way to write this sentence so that the novice reader is not mislead into believing that "an infinite continued fraction is an expression" has some mathematically rigorous meaning. (Re volumes: getting from 1+1=2 to \epsilon_0 was a whole lot easier, eh?) —Quantling (talk | contribs) 21:45, 8 December 2010 (UTC)
No, I would say, and do say, that our articles are - and must be - written in colloquial English; we are tertiary literature, intended to explain the secondary literature, not to reproduce it. It may be useful to tweak to something like "indefinitely expandible expression", but all these indicate the same thing. Septentrionalis PMAnderson 00:48, 9 December 2010 (UTC)

I do think that argument misunderstands the function of a lead section. It is not to provide the foundation of what follows, but to provide access to it. It also is a help to the reader who needs to know quickly whether the article found is the right one for the reference need. I'm all in favour of a bit of handwaving and heuristic in lead sections. As a figleaf one can say "roughly speaking". Charles Matthews (talk) 22:14, 8 December 2010 (UTC)

There is now a new article infinite expression. If it is to be kept, it is in desperate need of actual content: what is there now (infinite series, infinite products, and continued fractions) are very far from being adequate as a definition of an infinite expression, although they all make good examples of infinite expressions. —David Eppstein (talk) 16:58, 9 December 2010 (UTC)
The newest reference is from 1948 and one of the three is by Euler? If it was a historical article like versine, that might be OK. As things are I don't see why that article would be kept at AFD.
This is an example of what I call the "big number" phenomenon. After seeing several books say that some number or the other is "big", an editor makes an article "big number" that starts "In mathematics, a big number is ... ". I have seen at least one other example of this, although I don't remember it. Not every English phrase is a mathematical term, even if it is used in a mathematical context.
A more subtle problem is that there are examples of common phrases in math books that are used as if they are formally defined, when really they are not. Try to find a formal definition of an indeterminate form in a calculus book, for example. It is clear to me that we need an article on indeterminate forms, but far less obvious that we need an article on "infinite expressions". — Carl (CBM · talk) 17:24, 9 December 2010 (UTC)
I concur with Carl's observation of the "big number" phenomenon: good analogy. Unless there are actual references that treat the subject, rather than just by way of abuse of language, then this is original research by synthesis. The cited abstract might make a useful starting point for an actual article on the subject. In the mean time, it is probably worth stubbing the rest of the content until someone gets around to reading that paper. Sławomir Biały (talk) 16:01, 11 December 2010 (UTC)

Article assessment update

Here's a quick update on the article assessment / Wikipedia 1.0 progress this year. We have a little over 8,700 assessed articles at the moment, all of which have quality and priority rankings assigned. The latest offline release, Wikipedia 0.8, will come out soon, and includes approximately 900 mathematics articles. This selection of articles was chosen based primarily on article assessments.

I'm very interested in hearing from math editors if there are any changes to the assessment system that would be useful.

There is one visible change coming up soon, probably before January. For a couple years, VeblenBot updated tables and lists of mathematics articles grouped by quality assessment, priority assessment, and field. These are visible on Wikipedia:WikiProject Mathematics/Wikipedia 1.0 and its subpages. Unfortunately, we have outgrown that system; there are too many articles to make browsing them on wiki pages very effective, and the old system actually breaks down from time to time when pages get too long.

I will be migrating those pages to use a new system that uses a web tool on the toolserver to generate the article lists dynamically. This is the same system used to make the selections for release versions (the WP 1.0 bot). The summary tables will still be uploaded to the wiki. I have already written code to generate the tables using the new system; the new tables are visible at Wikipedia:WikiProject Mathematics/Wikipedia 1.0/Table/Sandbox and will be updated each day.

Please do let me know if you have suggestions or complaints about the system. There is no reason why we cannot change our assessment system over time to suit our needs as they change. — Carl (CBM · talk) 23:13, 12 December 2010 (UTC)

Morisita's overlap index

Morisita's overlap index has problems. I've commented at talk:Morisita's overlap index. Michael Hardy (talk) 01:56, 12 December 2010 (UTC)

It looks distinctly better now, although it would bear expansion. Michael Hardy (talk) 00:09, 15 December 2010 (UTC)

Complex projective space

This article really needs some work. It's had about 40 edits (excluding my half-dozen) in the last six years. It's been tagged with {{unreferenced}}, {{jargon}}, {{lead too short}} and {{expert}}. The last three are my tags. It's such a fundamental topic, but hasn't received any attention. It's totally impenetrable and goes against WP:NOT PAPERS and WP:JARGON. I outlined my concerns on the article's talk page. I've always focused on real geometry so I don't feel confident to make the changes. It needs humanising, it needs some enthusiasm and motivation; basically, it needs some life breathing into it. Any volunteers? Fly by Night (talk) 01:07, 14 December 2010 (UTC)

The article isn't breathtaking, but I don't think it's "so bad". The advice to avoid jargon has to be tempered with the advice to come to the point, especially in the lede. This article would at least be useful as a small reference for someone who has a basic background and. I have certainly seen much less useful articles.
In general, WP:NOT PAPERS has always been a little inaccurate. The idea that every term not known to the typical 10th grade student should be explained in every article is so impractical that we have never made an attempt to follow it. There is a kernel of truth in the advice there: an article here should not be written in the same way, or at the same level, as a research paper. But some editors would treat anything not covered in a typical grade school as "jargon".
I think an article on the complex projective plane will necessarily be more technical than an article on complex numbers. For example, the typical reader of complex number won't care whether the complex plane is simply connected, but the reader of an article on CPn is much more likely to care about such things. I don't think we should ignore that. — Carl (CBM · talk) 02:01, 14 December 2010 (UTC)
I agree that NOT PAPERS is inaccurate... at best. There's only so much that can be done. CRGreathouse (t | c) 03:48, 15 December 2010 (UTC)
There has been some work done in the last few days but I think the jargon issue is still there and it's a typical example of the kind of jargon problems seen in this project. Specifically, the lead and the first section after it contain a lot of language that only people who are already familiar with concept would be able to understand. For example there probably aren't too many people who know what a "line bundle" is who aren't already familiar with complex projective space. I'm thinking a typical reader who might be interested in the subject would an undergraduate math major, but such a reader would probably be confused and put off by all the unfamiliar terminology which appears so early in the article. The article needs to be organized, as should all articles, so that someone who knows just enough to have heard of the subject will understand the lead and, if possible, the first few sections.--RDBury (talk) 09:23, 15 December 2010 (UTC)
The problem isn't one of organization, it is of lack of content. There really isn't anything elementary in the article at all, so this already constrains the lead, and no amount of reordering of content is going to fix it. Also, as Carl points out, there is a certain amount of jargon that is unavoidable in the subject: the notion of a line bundle would be far too much to explain in situ (that's why god invented the Wikilink), yet it would be inappropriate not to mention the role of projective space as a classifying space. Sławomir Biały (talk) 12:42, 15 December 2010 (UTC)


As far as I know, there is no legitimate way of using \mbox within TeX in Wikipedia articles. Once upon a time, years ago, we didn't have \text and people used \mbox as a sort of substitute. The intended use of \mbox is to prevent line-breaks when TeX or LaTeX is used in the normal way, as opposed to the way it's used here. It's not intended to behave like \text, and in fact one routinely includes either text or math notation within \mbox when it's used in the normal way rather than the way it's used here.

There is a problem: In some contexts, \text and \mbox yield different results. Thus

 x_\mbox{initial}, \quad x_\text{initial} \,

(the fist uses \mbox; the second uses \text).

Somehow it seems as if \mbox is a persistent meme. How should we change that? Michael Hardy (talk) 00:16, 15 December 2010 (UTC)

It will be pretty hard. \text is an AMS extension to LaTeX, so a lot of people are in the habit of using \mbox for real papers. But we could do a scan of a database dump for articles that \mbox to see how many of those exist. — Carl (CBM · talk) 02:58, 15 December 2010 (UTC)
Why would they use \mbox rather than \rm or \mathrm or something like that in "real papers"? We're talking about writing text within a math environment. \mbox is used for things like this:
The only object satisfying that criterion \mbox{is $SL_2(\mathbb{R})$}.
The \mbox does not appear within a math environment, but has something within it that is in math mode. The purpose of \mbox is to prevent any line-break between the things within it. Michael Hardy (talk) 03:44, 15 December 2010 (UTC)
One easy thing to do is to remove the usage from Help:Displaying a formula. That's where many editors are look for best practice examples of formula writing.--RDBury (talk) 09:35, 15 December 2010 (UTC)
In plain TeX, the following is one way to return to text mode inside math, for example in a displayed formula:
$$a \mbox{ but not } b$$
Just using \rm doesn't work because it doesn't switch back to text mode (for example spaces are still eliminated between words in ${\rm cat dog}$. The LaTeX commands like \textrm do switch back to text mode, as does the AMS \text command, but I see people use \mbox for this often enough. — Carl (CBM · talk) 12:02, 15 December 2010 (UTC)
I'd like to point out that \text is not always a great solution either, and in the specific example above it would be preferable to use \mathrm as in
The difference is that if you are in an environment where the text is italic, then the contents of \text will also be in italic, while those of \mathrm will not. While this doesn't occur often on wikipedia, it could speak to why people haven't flocked to \text seeing as though it's not the silver bullet Michael Hardy claims it is. Are there occasions in wikipedia where one is in an italic environment (like in \begin{theorem} say)? RobHar (talk) 15:45, 15 December 2010 (UTC)
That's never happened for me when I use LaTeX on unix machines. The \text{} stuff is never italicized. Michael Hardy (talk) 21:17, 15 December 2010 (UTC)
The fact that \text adapts to the environment is actually is advertised purpose. It can be used in macros where it is not known ahead of time whether the invocation will be in subscript or not, and not known whether the environment will be italic, bold, normal, etc. For the subscript you mentioned, \textrm would also work. The \mathrm command stays in math mode, which has some consequences. First, it uses math-mode spacing, so \mathrm{foo: bar} is not the same as \textrm{foo: bar}. Second, text-mode commands like \textbullet won't work in \mathrm. — Carl (CBM · talk) 16:42, 15 December 2010 (UTC)
Oh yeah, I know \text is only doing what it's supposed to do, it's great for its purpose. I actually tried out \textrm before making my comment and it switched to italic when in an italicised environment. Is the behaviour of these things platform dependent? I use \text for text within math and \mathrm (or \operatorname) for math as in my example above. To have spaces in an subscript I guess you could just use \mathrm with an escaped space. RobHar (talk) 23:42, 15 December 2010 (UTC)
You're completely right: textrm doesn't change to upright shape, just the same font family as the body text. Thanks for pointing that out. It turns out there is "textnormal" that will do it. Re Michael Hardy: it should not be platform dependent, you can see the results in this file.

\textit{$a \text{ car} \textnormal{ dog} \textrm{ cat} \textup{ sheep}$}

\textsf{$a \text{ car} \textnormal{ dog} \textrm{ cat} \textup{ sheep }$}

— Carl (CBM · talk) 01:30, 16 December 2010 (UTC)

The solution depends on the problem. \text should be used for text, \operatorname should be used for operator names, and \mathrm should be used for other math material that needs to be upright (such as the "initial" label above). I can't think of any legitimate use of \mbox that would not involve features unsupported by texvc.—Emil J. 16:54, 15 December 2010 (UTC)

I think the first step is to fix Help:Displaying_a_formula so that it uses \text instead of \mbox in its examples. That might decrease the number of uses going forward. I did a scan of a database dump from May, and there are over 1700 articles the included the string '\mbox' when that dump was made [8]. I don't think it's worthwhile to go through and try to edit all of those, especially since in many cases the output wouldn't be any different. If there is any specific type of use of \mbox that's particularly bad and worth fixing, I can do another scan to look for articles that have it. — Carl (CBM · talk) 12:44, 15 December 2010 (UTC)

I'm am appalled to see that page condoning \mbox. I've changed it!

The "particularly bad" ones would include those where \mbox rather than \text or \mathrm is used in subscripts and superscripts, since \mbox makes the text much too big. Michael Hardy (talk) 21:24, 15 December 2010 (UTC)

Where's the article?

So the teacher "weights" the grade, meaning that 50% of the grade is from homework, 20% from the final exam, etc. Where's the article on this?

I found Weight (representation theory), but that looks unlikely. Is there another name for it? WhatamIdoing (talk) 20:15, 15 December 2010 (UTC)

Weighting (which is not linked from Weight (disambiguation) but should be) and Weighted mean cover this sort of weight, but neither specifically covers grading in education. Weighted average mark redirects to Grade (education), which doesn't seem to contain much of relevance. Algebraist 20:33, 15 December 2010 (UTC)
Thanks, that's what I wanted.
"Weighted grade" is a problem in the context of American education, because it means two different things: it can be weighting assignments to create an overall class grade, or weighting class grades to inflate the grade point average of students enrolled in harder classes (thus making it possible for a student to have a GPA of 4.2 on a scale that supposedly rises no higher than 4.0). WhatamIdoing (talk) 21:06, 15 December 2010 (UTC)
O, the days of weighted GPAs when I was in high school. It was possible (and still is in many places) for the valedictorian to graduate with close to a 6.0 GPA on a scale of 0 to 4. — Carl (CBM · talk) 01:35, 16 December 2010 (UTC)

Sicherman dice Images Requested/Needed

For the Sicherman dice article, I feel it is necessary for there to be the addition of two images:

  • An image of a pair of actual Sicherman dice
  • An image of the outlay of all the faces of both Sicherman dice

I was unable to find accurate or adequate images online that were free-use, so I would appreciate it if novel images or online images I did not see that meet Wikipedia standards could be provided for application via upload to Wikimedia Commons. If this is done, please contact me (Talk).

If I have incorrectly "posted" this talk section or have made an inappropriate request, please contact me (Talk).

I am a new user so please cut the appropriate amount of slack for any errors in format of conduct. QtheAllmighty, Talk. —Preceding undated comment added 23:58, 17 December 2010 (UTC).

operator (mathematics) and operation (mathematics)

I found that on Wikipedia what I suppose is a wide-spread confusion between the terms operator and operation. I don't have Ph.D. in mathematics, but so far in all discussions I had (abeit, mostly in Russian) there was a clear-cut distinction: the term operator refers to a map between vector spaces, and the term operation refers to a specific kind of mapping that gives a set an algebraic structure. From what I understand, the confusion might stem from the fact that the term operator denotes something analogous to operation in programming, see operator (programming). The end result was that the state of an article on operators (then just called operator) was awful: operators were discussed alongside operations, and the most input was, it seems, from programmers who didn't understand the difference between arithmetic operations and operators in programming, and were more than willing to post examples from various programming languages on what they percieved as their subject.

So I deleted most of the article and moved it to operator (mathematics), and now I'm digging through the links to the operator (which now redirects to disambiguation page), doing my best to disambiguate. I also put notes in both operator (mathematics) and operation (mathematics) in order to lessen the confusion.

I encountered several problems:

  • There is an article called operator (physics) that is currently dealing with applications of operators to classical and quantum mechanics. I find the title confusing when I'm trying to disambiguate links on physics-related pages. I proposed to move the article to Mathematical operators in physics, but I'm not quite sure.
  • I'm currently engaged in an argument with User:Patrick who thinks that operators are a special case of operations. I find the idea bewildering, but I still want to discuss it here just in case I'm wrong. See his talk page.

Finally, I want to urge you to collaborate on the series of articles which are really the basics of mathematics, but are sadly neglected and are currently in a very bad shape. These are, again:

I would also appreciate help with disambiguating links to operator (disambiguation). I will gladly do it myself (after all, I started all this mess), but still :)

Please, respond! — Kallikanzaridtalk 19:40, 20 December 2010 (UTC)

Operator and operation are intrinsically vague terms. The only purpose of a page on either is to just cover the most common usages, and then direct people to what they're really after. Right now, you've turned the operator (mathematics) page into a description of various finite- and infinite-dimensional linear operators, which seems like a duplication of stuff we already have. -- Walt Pohl (talk) 20:13, 20 December 2010 (UTC)
It's not entirely so because operators on vector spaces are more general than linear operators, and they are an important research topic AFAIK. What do you suggest? I wouldn't recommend rolling back everything, because the state of operator article before was horrible with little hope for improvement IMO. — Kallikanzaridtalk 20:36, 20 December 2010 (UTC)

5-D geometry Merges

Hi. An editor added a note (pasted below) at the Proposed Merges project page; doesn't seem controversial/hard to carry out so it's kinda outside the page's scope. The articles hadn't ever been tagged for merging or had any discussion about merging on their talkpages. I went ahead and placed merge (to "5-simplex") tags on the articles—I think that's what they had in mind. To be honest, given the subject's technical nature, if anywhere's likely to bring attention/feedback it's here.

Merge Bitruncated 5-simplex and Bitruncated hexateron. Hexateron redirects to 5-simplex. --

Thanks, Whitehorse1 21:32, 20 December 2010 (UTC).

The correct merge was Bitruncated 5-simplex and Bitruncated hexateron which has now been done.--Salix (talk): 22:18, 20 December 2010 (UTC)
Excellent, thanks Salix and Tomruen. --Whitehorse1 22:37, 20 December 2010 (UTC)

is internal set theory a constructivist theory?

Please comment here. Tkuvho (talk) 15:51, 22 December 2010 (UTC)

Math related FP nomination

There is a new nomination for a math related image (File:Theorem.gif) at Wikipedia:Featured picture candidates/Pythagorean Theorem Animation.--RDBury (talk) 13:40, 26 December 2010 (UTC)

It is garbage which was fortunately withdrawn from consideration. JRSpriggs (talk) 20:53, 26 December 2010 (UTC)

Pseudolinearity and pseudolinear functions

I've created a placeholder redirect (that can be renamed over) at pseudolinearity, which is probably where this article is best sited, with some refactoring of the definition in the introduction of course. I've also created a redirect at η-pseudolinearity. Opinions on the title are welcome. Observe too the redlink at invexity, another article that we didn't even know that we didn't have. ☺ Uncle G (talk) 23:16, 22 December 2010 (UTC)

  • I dislike abstract nouns being used as article names since it's really just a variation on using adjectives as article names. Adjectives in mathematics, and the corresponding abstract nouns, tend to get reused and can mean very different things depending on what they're applied to, so they are usually a poor indicator of what the subject is. An example would be "normality"; would that be about normal distributions or normal spaces?--RDBury (talk) 17:48, 23 December 2010 (UTC)
    • You should be arguing that with all of the mathematicians that use "pseudolinearity" in the titles of their journal articles, book chapters, and whatnot. ☺ Uncle G (talk) 22:56, 23 December 2010 (UTC)
      • I don't really object in those cases, they have context built in. Wikipedia is different in that you have to assume people may come to an article dropped in by parachute, so it's helpful to put a bit of extra information in the title to help them locate where they are.--RDBury (talk) 11:26, 25 December 2010 (UTC)

I've recently started the article pseudoconvex function as a more appropriate subject (for essentially this same content). The article is just a few hours old, but already is the subject of an apparently non-mathematical editor with a hardline view of WP:NOT#TEXTBOOK. I'd appreciate any help with the article contents to make it more reader-friendly. Sławomir Biały (talk) 17:00, 25 December 2010 (UTC)

The editor in question apparently believes that Wikipedia should not cover advanced mathematics topics. Comments at Talk:Pseudoconvex function would be appreciated. Sławomir Biały (talk) 18:01, 25 December 2010 (UTC)
I added some references and a short discussion of linear fractional programming in the pseudoconvex function article. SB's addition is greatly appreciated (by me, at least). Kiefer.Wolfowitz (talk) 10:35, 29 December 2010 (UTC)

pi in the sky, plane, or space?

Is the circle in the plane or in space? Comment at talk:pi. Tkuvho (talk) 20:15, 27 December 2010 (UTC)

Mathematics Genealogy Project

In the last month, editor Edstat renewed his campaign against the Mathematics Genealogy Project: c.f., the October notice to this WP project (by David Eppstein). I removed statements in the MGP article that seemed to violate (imho) original-research (e.g. synthesis) and undue-weight prohibitions. As in his previous campaign, Edstate contributed several cases of selective & truncated quotation, imho.

As a quality-check on my edits and possibly on Edstat's editing, other editors should review the MGP article.

Sincerely, Kiefer.Wolfowitz (talk) 10:43, 29 December 2010 (UTC)

Math-related FA on the Main Page tomorrow

Hi, I just wanted to alert you all that the Laplace–Runge–Lenz vector will be appearing on the Main Page tomorrow, that is, on December 30th. Strictly speaking, it's more of a physics article than a math article; but it does appear on our list of WPM Featured Articles. I'll do my best to watch over it tomorrow, but I'll also be pretty busy in real life, so I'd appreciate any extra help from my friends and colleagues here. Thank you very much! :) Willow (talk) 18:28, 29 December 2010 (UTC)