I've discovered how to break it badly :)
I just re-enabled it and trued changing the rendering to MathML, changing two things in the popup you get on right-clicking on a formula. And now:
It looks like just a missing file, though it may be something you should advise testers about if it's not supported. I can just go back to PNG rendering for now, so won't try doing anything to fix it myself.-- JohnBlackburne words deeds 03:15, 27 December 2010 (UTC)
The URL you give is sourceforge, and the page states that it is no longer up-to-date. Maybe it's better to post a newer URL. Or is this Mathjax an old version that requires old fonts? -- Netheril96 ( talk) 03:42, 22 March 2011 (UTC)
Aside from Help talk:Displaying a formula#Formulas as SVG? and your own userspace, is there any place people can know of your MathJax extension? It needs and deserves to be more popular. -- Netheril96 ( talk) 08:20, 23 March 2011 (UTC)
It seems whenever a formula is followed by text or citation, it will be automatically reduced to text style regardless of the colon before it (or lack thereof). For example,
-- Netheril96 ( talk) 13:47, 23 March 2011 (UTC)
where r is the position vector.
Another bug, see Rotation_representation_(mathematics)#Rotation_matrix.-- Netheril96 ( talk) 03:14, 24 March 2011 (UTC)
Every time I browse photon with Chrome, it crashes when math is processed 83%.-- Netheril96 ( talk) 01:59, 28 March 2011 (UTC)
On Curvilinear coordinates, MathJax processing stops at 31%. I refreshed several times and bypassed cache but still it wouldn't finish. Using Firefox.-- Netheril96 ( talk) 10:48, 29 March 2011 (UTC)
Thanks for the great work of porting MathJax to Wikipedia!
When browsing certain pages I found that MathJax could get stuck at loading [MathJax]/jax/output/HTML-CSS/fonts/TeX/Main/Bold/SpacingModLetters.js
followed by a message of the file failing to load. Can you confirm whether the file is missing or not?
Kxx (
talk |
contribs)
05:22, 3 May 2011 (UTC)
Looking at Photon I noticed a strange bit of maths formatting at the bottom of the page. Tracing down the problem I reduced it down to User:Salix alba/MathJaxBug. In my browser Safari the two equations which should display in the second column of the references actually appear below the reference block. The page crashes Chrome.-- Salix ( talk): 11:26, 5 May 2011 (UTC)
The code I'm using is
<script src="../MathJax/MathJax.js">
MathJax.Hub.Config({
jax: "input/TeX","output/HTML-CSS"],
extensions: "tex2jax.js","MathMenu.js","MathZoom.js"],
TeX: {
extensions: "AMSmath.js","AMSsymbols.js","noErrors.js","noUndefined.js"
}
});
</script>
<div style="column-count: 2; -moz-column-count: 2; -webkit-column-count: 2; list-style-type: decimal;">
<ol class="references">
<li id="cite_note-0">A \(a^1\)</li>
<li id="cite_note-0">B \(b^2\)</li>
<li id="cite_note-0">C \(c^3\)</li>
<li id="cite_note-0">D \(d^4\)</li>
<li id="cite_note-0">E \(e^5\)</li>
<li id="cite_note-0">F \(f^6\)</li>
</ol>
<ul>
</div>
When there is bold-italic font in math, mathJax will load Web-Font bolditalic, like this :, which takes a long time to load in math intensive articles. But I have already installed all MathJax fonts, including MathJax_Math-BoldItalic. Why doesn't it use local fonts? -- Netheril96 ( talk) 13:30, 25 May 2011 (UTC)
Actually, it is a developer on MathJax user group who made a mistake. He at first told me that the next release were going to fix the MathML problem, that is, missing space between functions like sin, cos, ln and their arguments. Just now he says that only a fork [1] fixes that; planned release doesn't include this. -- Netheril96 ( talk) 02:19, 28 May 2011 (UTC)
After days I still can't stand the long waiting period of "Loading Web-Font". Are you sure your Firefox 4.0.1 doesn't have such a problem? I even tried a clean Firefox 5 (beta 3) without any addons, but the problem is still there.-- Netheril96 ( talk) 03:17, 4 June 2011 (UTC)
It's good to hear that turning off Direct2D support in Firefox finally solved the problem. For any reader running into the same issue, you can do that by unchecking hardware acceleration under "Preferences" menu -> "Advanced" tab. Alternatively, there are a few switches in "about:config" dealing with "gfx" rendering issues, such as "gfx.direct2d.disabled".
For your desire of having "bolder" fonts by default, that's hardly doable because there are only two type faces, "normal" and "bold". So if you set "normal" fonts "bolder" they will become "bold" but the "bold" symbols won't become darker. The best you could do is to simulate darker characters by rendering them twice. For example, adding the following line to your vector.js will print the second run with a vertical offset of 1 pixel:
Don't think you will like the result, though. Nageh ( talk) 20:15, 7 June 2011 (UTC)
\tfrac: ['\\dfrac{#1}{#2}',1,2]
A discussion on mathJax is current underway at Wikipedia_talk:WikiProject Mathematics#MathJax.-- Salix ( talk): 07:30, 25 June 2011 (UTC)
Hello there. Firstly, thanks a lot for all of your work involving MathJax. One problem that I have notice is with subscript square bracket: . The final bracket is very faint and is almost invisible. — Fly by Night ( talk) 23:01, 25 June 2011 (UTC)
Another thing that needs attention:
Michael Hardy ( talk) 04:01, 27 June 2011 (UTC)
Another similar bug:
Michael Hardy ( talk) 23:15, 28 June 2011 (UTC)
\underline \mathbb Z_X
and \hat \boldsymbol \theta
fail in the "amsart" documentclass. As MathJax does not implement whole documentclasses there is no way of telling what would be the "right" interpretation of parameter precedence. The third problem is that \hat \boldsymbol \theta
and \hat {\boldsymbol \theta}
are typographically and semantically wrong. You do not want a bold-formatted vector that is modified with a normal-formatted hat on it, but you want a bold-formatted vector whose notation includes a (bold) hat on it. Concerning \underline \mathbb Z_X
, there may be a way to special-case \mathbb
... I'll have to investigate into this.For those using MathML to render output, I have fixed a bug concerning incorrect rendering of "special" font styles such as blackboard bold. Nageh ( talk) 18:46, 7 July 2011 (UTC)
Here's the code:
and now I put it within math tags:
The way this now appears in my browser window is that I no longer see this
but instead I see this
So the bug that was there earlier is gone, but it's still not putting the 45 points of extra vertical space between lines. There are times when that feature is needed, and it currently appears in MANY Wikipedia articles. Michael Hardy ( talk) 19:24, 30 June 2011 (UTC)
Since I started viewing things on Wikipedia via mathJax, the blackboard-bold C below appears much smaller:
Does mathJax do that somehow? Michael Hardy ( talk) 18:02, 5 July 2011 (UTC)
This one should make you happy, Michael. :) One of the MathJax developers (dpvc) has committed a patch for this issue, and I have just integrated it into my user script. Nageh ( talk) 12:58, 4 August 2011 (UTC)
In the references section to Erdős–Straus conjecture ( this version) there is a title of a paper by Tao with \frac in some math. When I view it in Chrome (on OS X) I see only " = + + "; that is, the stuff inside the fractions is missing. Maybe this has something to do with being in a small font and two column references format? Anyway it seems to be a bug. In the previous revision the math had a \scriptstyle in it but I took it out, thinking that might help; nothing changed, though. — David Eppstein ( talk) 20:30, 7 July 2011 (UTC)
I notified the mathJax developers of the existence of this present page and they seem to be looking it over. Michael Hardy ( talk) 22:55, 21 July 2011 (UTC)
Annoying as it is, I still like mathJax better than the default non-scalable, non-customizable image mode. Bugs certainly need fixing, but you should not wait until every aspect is impeccable to start using or promoting something.— Netheril96 ( talk) 05:34, 25 July 2011 (UTC)
Michael, that bug where you would actually see [10pt] has long been fixed. This was truly a show stopper, but missing the intended vertical space is more a flaw than a bug. Nageh ( talk) 08:18, 25 July 2011 (UTC)
Hi Nageh, thanks for your efforts on this. I hope this is the appropriate place to post a bug (?) report.
I followed the installation instructions. Now when I load a page with math on it mathjax tries to load, but stalls with this message:
"Loading [MathJax]/jax/output/HTML-CSS/fonts/STIX/fontdata.js"
After maybe 10 seconds mathjax gives up loading the file, gives a file-not-found sort of error, and then is stuck with the message
"Processing math: 0%"
I'm using the Vector skin, and have this in the common.js skin file:
mathJax={}; mathJax.fontDir=" http://cdn.mathjax.org/mathjax/latest/fonts"; importScript('User:Nageh/mathJax.js');
I tried it also without the first line above (starting with "mathJax") with the same result.
If I disable the STIX fonts on my computer, I do not have the problem. (I'm using a mac with OSX 10.7 (Lion) which comes with STIX fonts installed.)
Thanks Nealeyoung ( talk) 04:21, 31 July 2011 (UTC)
I've just noticed that while using mathJax,
does not look the same as
They both gave identical results before I started using mathJax. Michael Hardy ( talk) 12:57, 6 August 2011 (UTC)
IE has nasty performance issues with .offset{Width,Height,Left}, which MathJax heavily relies on. There is a ticket at Microsoft Connect that complains about this. MathJax updates the DOM immediately after each formula is typeset, typically introducing a lot of new nodes, and causes a full-page re-layout. Such quadratic behavior, exacerbated by slow accesses to .offset{Width,Height,Left}, renders MathJax painfully slow on pages like " List of trigonometric identities" which contain huge numbers of formulae.
A workaround is to prevent IE from frequent re-layouts. My approach is to let MediaWiki generate PNG as usual, do something like
$('img.tex').each(function() {
var span = document.createElement('span');
span.className = 'tex';
span.textContent = this.alt;
this.appendChild(span);
});
then register a listener for the "End Typeset" signal
MathJax.Hub.Register.StartupHook('End Typeset', function() {
$('img.tex').each(function(){
var parent = this.parentNode;
parent.insertBefore(this.children1], this);
parent.replaceChild(this.children1], this);
});
});
Since children of IMG elements do not participate in layout, MathJax's DOM updates have minimal impact on performance. This enables IE to typeset noticeably faster while allowing all formulae to be visible throughout.
Perhaps similar logic can be baked into the MathJax startup script. Kxx ( talk | contribs) 23:21, 9 August 2011 (UTC)
I recall the MathJax main developers having discussed such issues before though I haven't checked whether any relevant code has been added to the code repository. There are related discussions at the MathJax developers mailing list, e.g., [3]. You might want to bring this up at that mailing list or at the user mailing list as it is really something that should be integrated into the MathJax core code rather than my script only. Thanks, Nageh ( talk) 11:25, 10 August 2011 (UTC)
In Expectation–maximization algorithm, after the heading "Alternative description", the two lines of TeX after "expectation step" and "maximization step" are not getting rendered when I view them while logged in (and using MathJax. All I see is the TeX code. What's going on? Michael Hardy ( talk) 19:51, 5 October 2011 (UTC)
At https://bugzilla.wikimedia.org/process_bug.cgi I put in a feature request in February. This morning I looked at User:Michael Hardy/scratchwork and found that \koppa, \stigma, and \coppa within TeX are finally getting rendered! I was delighted to see this. But now it turns out that they work properly only when I not logged in! That makes me suspect the problem is with MathJax.
I need these in order to continue work on Ptolemy's table of chords. And there are probably other articles in which they should be used. Michael Hardy ( talk) 20:16, 5 October 2011 (UTC)
When I'm logged in and using MathJax, this table doesn't appear the same way it appears when I'm not logged in. I intended that between the first and second columns, and again between the fourth and fifth columns, two vertical lines would be visible. That is in fact how it appears when I'm not logged in. But when I'm logged in, and hence using MathJax, I see only a single vertical line there. The comprehensibility of the table depends on those two vertical lines being marked as different from the others.
Another difference is that when I'm logged in and using MathJax, I see vertical lines at the left and right edges of the table, and those are not there at all when I'm not logged in. Michael Hardy ( talk) 02:54, 14 October 2011 (UTC)
This looks very good when I'm not logged in, but when I'm logged in (and hence using mathJax), I get neither an error message nor the intended image; rather I just get TeX code, all on one long line, so that I have to scroll a long distance to the right to see the end of it. What's going on? Michael Hardy ( talk) 14:57, 15 October 2011 (UTC)
\overset\text{`}
, which MathJax interprets as \overset{\text}{`}
– change it to \overset{\text{`}}
, as I have done above.
Nageh (
talk)
15:30, 15 October 2011 (UTC)
Look at this:
The code between the math tags is as follows:
What appears on my browser is what I'd expect if I'd written the following:
Let's try the latter code:
This could lead to grave misunderstandings: the reader would think that what's intended is the greatest integer less than or equal to X. Michael Hardy ( talk) 14:44, 17 November 2011 (UTC)
One the N-sphere page the function \arccot is displaying in red. I guess that this is a function defined in texvc and not in MathJax. It would be possible to replace the \arccot with \operatorname{arccot} but should we be adding workarounds when texvc is still the default renderer.-- Salix ( talk): 08:29, 9 January 2012 (UTC)
MathJax crams some things together that currently has a space between. For instance in
MathJax puts no space between the second lim and x Dmcq ( talk) 17:35, 29 February 2012 (UTC)
It seems like MathJax 2.0 has indeed resolved this issue. I have just upgraded to the new release. As always, report any bugs you run into on this talk page, and thanks for testing! Nageh ( talk) 16:25, 1 March 2012 (UTC)
If an maths expression is positioned at the start of a new line because it won't fit on the last and then hover over it, e.g. try the magnified image appears at the right at the end of the last line. Dmcq ( talk) 22:50, 1 March 2012 (UTC)
I the current scheme if one makes a error in the tex the whole expresssion is shown in red in the preview. However I noticed that none of the expressions was output when I had an error and was using MathJax. For instance in
If you take the nowiki tags off the end none of the them show up whereas only the second one has an error with two back slashes. This can be a bit disconcerting needing a binary chop if the error isn't obvious. Dmcq ( talk) 11:37, 7 March 2012 (UTC)
Mathjax 2.0 changes behavior such that MJ 2.0 no longer makes use of firefox mathml by default. I am only mentioning this to you in case you want to update the firefox mathml discussion on the mathjax installation page.
http://www.mathjax.org/2012/03/02/news/mathjax-2-0-and-the-default-rendering-in-firefox/
DouglasCalvert ( talk) 03:38, 11 March 2012 (UTC)
I would like to use mathjax on the Discrete Probability page from wikibooks. Is there an easy way to do this?
http://en.wikibooks.org/wiki/High_School_Mathematics_Extensions/Discrete_Probability
DouglasCalvert ( talk) 04:14, 11 March 2012 (UTC)
Any updates on when mathJax will be ready to add to wikibooks? Creating a wikibook here:
http://en.wikibooks.org/wiki/Parallel_Spectral_Numerical_Methods
which could desperately use more LaTeX math features, in particular automatic equation numbering. Any answers would be particularly welcome in the discussion here:
http://en.wikibooks.org/wiki/Wikibooks:Reading_room/Proposals
Benson Muite ( talk) 12:39, 18 November 2012 (UTC)
Hi, Nageh. Have you seen this script from MathJax which says it takes the images, back-renders them into the tex, and calls MathJax. I'm no programmer (nor do I play one on the radio), so is this helpful to you or us? Thanks, and thanks for writing your script! -- Avi ( talk) 20:08, 12 March 2012 (UTC)
Mathjax 2.0 added a new rendering method: SVG. It looks clearer in my opinion. Any chance that you may make it functional on Wikipedia? — Preceding unsigned comment added by Netheril96 ( talk • contribs)
The following displays with the old scheme but not with MathJax
Cheers Dmcq ( talk) 14:13, 23 March 2012 (UTC)
\dot{\vec u}
).
Nageh (
talk)
15:31, 23 March 2012 (UTC)How can I use the locally-installed STIX font set? — Edokter ( talk) — 10:19, 1 April 2012 (UTC)
How can I check is mathJax loaded on a given Wikipedia page? Could you implement some "debug mode" with extra information such as "some required components loaded"… "nn chunks of math found"… "nn chunks processed"… "nn chunks contain errors"? Incnis Mrsi ( talk) 18:49, 2 April 2012 (UTC)
javascript:alert(typeof(mathJax)!="undefined"&&mathJax.loaded)
into your address bar of the current browser window to check whether MathJax has been loaded. Once MathJax has been loaded, you can right-click on any maths formula and invoke the About MathJax entry in the menu, which will give you a list of loaded configuration items.Where is (will) the base CSS for MathJax (be) located? There are some padding issues that need to be fixed, such as the excesive vertical padding around stand-alone formulas. — Edokter ( talk) — 13:10, 15 April 2012 (UTC)
<dd>
tags that they have some top and bottom margin (0.2em and 0.5em). MathJax (or rather, the .MathJax_Display class) uses top and bottom margin of 1em by default... that is very different from LaTeX.\displaystyle
.:<math>
, that is, started with a colon. What happened was the mathJax was literally interpreting the maths as to be rendered in scriptstyle font, which resulted in this tiny, illegible font. I have resolved this issue now and will update the mathJax documentation. I hope this resolves one of your issues.I'm just trying the "quick test" of your version of MathJax, since the standard (experimental) version is too obviously broken. I see that the \scripstyle
being ignored makes the integration region expression too large in the {{
oiint}} and {{
oiiint}} templates. Adding braces as stated does not appear to fix this. How to get the desired subscript size? —
Quondum
☏
20:34, 13 May 2012 (UTC)
While it looks like the experimental extension is using the Tex-AMS-MML_HTMLorMML.js configuration file, I would like to be able to use Tex-MML-AM_HTMLorMML.js so that I can also make use of asciimath2jax to display ASCIIMathML You provide the ability to tweak the Hub settings, but I think not so I can change the config file. What are your thoughts on this? Net-buoy ( talk) 23:58, 17 May 2012 (UTC)
Hello, Nageh.
There is now an experimental option in Wiki preferences to enable MathJax as the math viewer. Is this an implementation of your script, or something different, and do you recommend removing your script from JS now that there is this option? Thank you! -- Avi ( talk) 13:52, 30 May 2012 (UTC)
Whenever I edit an article with math in it, the Show preview button leaves the math portions as plain TeX in black (e.g. $ \mathbf{v} = (v_x,v_y,v_z). $). The math displays fine when I save the edit, though. Is this normal? My math appearance preference is set to Leave it as TeX (for text browsers) (although MathJax has the same problem) and my
User:Tschwenn/vector.js contains only one line: importScript('User:Nageh/mathJax.js');
--
TSchwenn (
talk)
21:59, 30 May 2012 (UTC)
javascript:alert(typeof(mathJax)!="undefined"&&mathJax.loaded)
into your address bar when you are editing an article and tell me whether it returns "true" or "false"?
Nageh (
talk)
22:23, 30 May 2012 (UTC)javascript:alert(typeof(mathJax)!="undefined"&&mathJax.loaded))
(close parentheses)? Putting either into the Firefox address bar and pressing enter does nothing at all, and in internet explorer it just redirects to the non-existent
http://www.javascript:alert..
. I tried suppressing the www.\%s.com auto-correct rule in my registry, but then IE just sent me to a Bing search for the command. On my other Windows 7 PC, both Chrome and Firefox do nothing with the command either. --
TSchwenn (
talk)
04:06, 31 May 2012 (UTC)alert(typeof(mathJax)!="undefined"&&mathJax.loaded)
and it should return either "true" or "false" (hopefully).
Nageh (
talk)
09:45, 31 May 2012 (UTC)false
everywhere, including after clicking Show preview. I have Firefox 12.0. --
TSchwenn (
talk)
12:49, 31 May 2012 (UTC)alert(MathJax.Hub.config.config)
gives, and whether mathJax.Config();
or MathJax.Hub.Startup.onload();
has any effect when you are previewing a page with unrendered math?
Nageh (
talk)
20:13, 31 May 2012 (UTC)TeX-AMS-texvc_HTML.js
. The other two commands do this in the web console: [16:30:28.821] mathJax.Config(); [16:30:28.834] undefined -- [16:30:58.644] MathJax.Hub.Startup.onload(); [16:30:58.657] undefined
. There are no apparent effects on the page. --
TSchwenn (
talk)
21:36, 31 May 2012 (UTC)MathJax.Hub.Update();
, MathJax.Hub.Reprocess();
, MathJax.Hub.Rerender();
do anything?
Nageh (
talk)
17:07, 1 June 2012 (UTC)(function anonymous() {return arguments.callee.execute.apply(arguments.callee, arguments);})
--
TSchwenn (
talk)
19:44, 1 June 2012 (UTC)Guys, the MathJax CDN serving the web fonts is currently experiencing some problem. This means that everyone who has included mathJax={}; mathJax.fontDir="
http://cdn.mathjax.org/mathjax/latest/fonts";
in his/her vector.js or common.js file will currently not be able to see any rendered math but only plain TeX code. I will inquire the MathJax folks about it.
Nageh (
talk)
18:14, 31 May 2012 (UTC)
mathJax={}; mathJax.fontDir="
https://c328740.ssl.cf1.rackcdn.com/mathjax/latest/fonts";
and it still seems to work fine. --
Avi (
talk)
18:51, 31 May 2012 (UTC)
Issues seems to be resolved. Nageh ( talk) 20:05, 31 May 2012 (UTC)
If I select the SVG option the browser outputs TeX, and I don't know how to get back to showing the formulae again with MathJax to change the option. Any suggestions? Thanks. I'm using Google Chrome with your javascript. I notice a box at the bottom saying 'Loading [MathJax]/jax/output/SVG/config.js' and then it puts out a long line with Failed to load and a full name with that name at the end. Dmcq ( talk) 23:23, 1 June 2012 (UTC)
In Squared triangular number, two of the references have mathematical formulae in them (with binomial coefficients, so impossible to rewrite as html). When I view them using Nageh/mathjax, I see a spurious italicized "< spanstyle="display:none;">< /span >" at the end of the reference. For example (with some minor changes to the TeX markup that don't affect the problem):
{{
cite journal}}
: CS1 maint: multiple names: authors list (
link)displays as something approximately like
(where I've replaced the binomial coefficient with the letter X). Oh, and it's a math italic font, not the text italic, and the angle brackets in the span don't look quite the same. Looking at the html source code generated here, I see <span class="tex"><span style="display: none;"> </span></span> which apparently is what ends up getting displayed in this way via mathjax attempting to interpret the html code as math.   is a no-break space character. Any ideas why this is happening, and how to fix and/or work around this? — David Eppstein ( talk) 21:51, 13 June 2012 (UTC)
<span class="tex"><span style="display: none;"> </span></span>
. I invite you to report this bug at
bugzilla. I am not going to do this, and other than this response I decided to be no more involved in Wikipedia activities because of the demonstrative arrogance of the WMF and its developers. It's sad, but obviously editor retention is not an issue to the WMF when it comes to established editors. HTH,
Nageh (
talk)
13:33, 14 June 2012 (UTC)Well, it doesn't happen when I link the title by hand: "Two quick combinatorial proofs of ". So maybe it is something in the cite template that is triggering this rather than in mediawiki? — David Eppstein ( talk) 16:10, 14 June 2012 (UTC)
This strongly hints to a bug in MediaWiki. There aren't supposed to be any <span class="tex"></span>
tags in the HTML source when the PNG rendering option is chosen, yet it appears in your example.
Nageh (
talk)
21:52, 14 June 2012 (UTC)
I'm finding MathJax incredibily slow (on each page load the browser takes ages loading fonts and .js files) – so slow that I set my preferences to PNG display for this reason alone. Any idea what this might be about? — Quondum ☏ 18:52, 16 June 2012 (UTC)
Would installing the STIX fonts locally help? -- Avi ( talk) 17:01, 17 June 2012 (UTC)
MathJax has this nice feature where clicking on a formula shows you a zoomed view of it. But when I click on the array in Wythoff array (in the Monobook format for Wikipedia as is my default preference), the zoomed view of the array is obscured by the Wikipedia search form, which comes up on top of the array. Is this normal? — David Eppstein ( talk) 02:18, 7 July 2012 (UTC)
<div id="p-search">
element, which prevents MathJax from overriding the layering order because it operates in a different
stacking context. One solution to solve this is to ask the MediaWiki devs to omit the z-index from the CSS... I don't see why a manual z-index would be needed at all. Another solution... well, looking at the MathJax sources it seems that the positioning element for the popup gets created as a child to the maths's immediate parent rather than as a child to the HTML root element... I guess that changing this could resolve the issue as well, but I'd need to inquire the MathJax folks on this. HTH for now.
Nageh (
talk)
10:56, 7 July 2012 (UTC)MathJax does not load in IE10 because it attempts to access document.namespaces
, which is not present in IE10. A workaround to manually create a dummy:
if (!document.namespaces) document.namespaces = { add: function() {} };
before MathJax is loaded. Kxx ( talk | contribs) 07:00, 28 August 2012 (UTC)
document.namespaces
is tested. Code that touches document.namespaces
should be removed/replaced (document.namespaces
itself is a nonstandard extension).
Kxx (
talk |
contribs)
15:46, 28 August 2012 (UTC)document.namespaces[e]
.
Kxx (
talk |
contribs)
23:00, 11 September 2012 (UTC)I switched to https mode a while ago, and the mathJax script is blocked by Chrome for it is not transferred in https. How should I modify the line importScript('User:Nageh/mathJax.js') so that the script is transferred from https://en.wikipedia.org/User:Nageh/mathJax.js instead?- Netheril96 ( talk) 14:36, 31 August 2012 (UTC)
I'm noticing a bug in MathJax, whereby the a is bolded italic, and the b is unbolded italic in <math>\mathbf{ab}</math>
:
In PNG it is it is the same as for <math>\mathbf{ab}</math>
(bolded roman), here also for MathJax:
Any ideas? — Quondum 06:34, 10 September 2012 (UTC)
<math>\mathbf{ab}</math>
is bold italic in MathJax, and bold roman in PNG, which is of course still a problem. I'm not sure whether \bold
is intended to be italic, but they should not be different. —
Quondum
13:59, 10 September 2012 (UTC)While I'm logged in, and looking at socle (mathematics), I see this line and it looks fine:
But looking at the same page when I'm not logged in, I see nothing but error messages where that line should be. What's going on? Michael Hardy ( talk) 16:39, 21 November 2012 (UTC)
I reported this at Wikipedia_talk:WikiProject_Mathematics#Math_Jax_problem
I switched over to Math Jax. On my browser (Windows IE 9, version 9.0.8112.16412, update 9.0.13) the formula
does not render correctly (the \not\equiv should be ≡ with a line through it, but is just a space on my browser)
Thanks.
And thanks for all the work on Math Jax; it looks much better than what I was getting before!
- Virginia-American ( talk) 17:04, 3 March 2013 (UTC)
Is this intended behaviour? It's not uncommon to see the markup, but then it disappears, all I see is horizontal lines where the fractions should be when the MathJax status goes away. Reloading doesn't seem to fix it, but purging does - implying it's server-side. Not a big deal for someone who's heard of purging - not so good for those who haven't. ChrisHodgesUK ( talk) 12:20, 24 June 2013 (UTC)
The page uses the typewriter font, but apparently [MathJax]/jax/output/HTML-CSS/fonts/TeX/Typewriter/Regular/Main.js is missing. Kxx ( talk | contribs) 07:47, 21 August 2013 (UTC)
User:Nageh/mathJax.js contains mathJax.version = "0.2.3";
, while
User:Nageh/mathJax/MathJax.js contains if(!MathJax.Hub){MathJax.version="2.2";
...}
. The latter is what is displayed when I right-click on some math and choose "About MathJax," so I assume it's correct. Why do these version numbers not agree, and does it matter for correct operation? --
SoledadKabocha (
talk)
05:22, 15 February 2014 (UTC)
Hi, it's not obvious to me what, if any, relationship there is between your user script and the experimental MathJax rendering option in Preferences, and the mw:Extension:MathJax. Does Wikipedia use that extension? I'm guessing no, but wanted to make sure. Also, where does discussion of math rendering on Wikipedia take place? I am wondering if MathJax will ever be turned on by default, and have no idea where to ask the question. Thanks! Klortho ( talk) 02:45, 24 March 2014 (UTC)
I've discovered how to break it badly :)
I just re-enabled it and trued changing the rendering to MathML, changing two things in the popup you get on right-clicking on a formula. And now:
It looks like just a missing file, though it may be something you should advise testers about if it's not supported. I can just go back to PNG rendering for now, so won't try doing anything to fix it myself.-- JohnBlackburne words deeds 03:15, 27 December 2010 (UTC)
The URL you give is sourceforge, and the page states that it is no longer up-to-date. Maybe it's better to post a newer URL. Or is this Mathjax an old version that requires old fonts? -- Netheril96 ( talk) 03:42, 22 March 2011 (UTC)
Aside from Help talk:Displaying a formula#Formulas as SVG? and your own userspace, is there any place people can know of your MathJax extension? It needs and deserves to be more popular. -- Netheril96 ( talk) 08:20, 23 March 2011 (UTC)
It seems whenever a formula is followed by text or citation, it will be automatically reduced to text style regardless of the colon before it (or lack thereof). For example,
-- Netheril96 ( talk) 13:47, 23 March 2011 (UTC)
where r is the position vector.
Another bug, see Rotation_representation_(mathematics)#Rotation_matrix.-- Netheril96 ( talk) 03:14, 24 March 2011 (UTC)
Every time I browse photon with Chrome, it crashes when math is processed 83%.-- Netheril96 ( talk) 01:59, 28 March 2011 (UTC)
On Curvilinear coordinates, MathJax processing stops at 31%. I refreshed several times and bypassed cache but still it wouldn't finish. Using Firefox.-- Netheril96 ( talk) 10:48, 29 March 2011 (UTC)
Thanks for the great work of porting MathJax to Wikipedia!
When browsing certain pages I found that MathJax could get stuck at loading [MathJax]/jax/output/HTML-CSS/fonts/TeX/Main/Bold/SpacingModLetters.js
followed by a message of the file failing to load. Can you confirm whether the file is missing or not?
Kxx (
talk |
contribs)
05:22, 3 May 2011 (UTC)
Looking at Photon I noticed a strange bit of maths formatting at the bottom of the page. Tracing down the problem I reduced it down to User:Salix alba/MathJaxBug. In my browser Safari the two equations which should display in the second column of the references actually appear below the reference block. The page crashes Chrome.-- Salix ( talk): 11:26, 5 May 2011 (UTC)
The code I'm using is
<script src="../MathJax/MathJax.js">
MathJax.Hub.Config({
jax: "input/TeX","output/HTML-CSS"],
extensions: "tex2jax.js","MathMenu.js","MathZoom.js"],
TeX: {
extensions: "AMSmath.js","AMSsymbols.js","noErrors.js","noUndefined.js"
}
});
</script>
<div style="column-count: 2; -moz-column-count: 2; -webkit-column-count: 2; list-style-type: decimal;">
<ol class="references">
<li id="cite_note-0">A \(a^1\)</li>
<li id="cite_note-0">B \(b^2\)</li>
<li id="cite_note-0">C \(c^3\)</li>
<li id="cite_note-0">D \(d^4\)</li>
<li id="cite_note-0">E \(e^5\)</li>
<li id="cite_note-0">F \(f^6\)</li>
</ol>
<ul>
</div>
When there is bold-italic font in math, mathJax will load Web-Font bolditalic, like this :, which takes a long time to load in math intensive articles. But I have already installed all MathJax fonts, including MathJax_Math-BoldItalic. Why doesn't it use local fonts? -- Netheril96 ( talk) 13:30, 25 May 2011 (UTC)
Actually, it is a developer on MathJax user group who made a mistake. He at first told me that the next release were going to fix the MathML problem, that is, missing space between functions like sin, cos, ln and their arguments. Just now he says that only a fork [1] fixes that; planned release doesn't include this. -- Netheril96 ( talk) 02:19, 28 May 2011 (UTC)
After days I still can't stand the long waiting period of "Loading Web-Font". Are you sure your Firefox 4.0.1 doesn't have such a problem? I even tried a clean Firefox 5 (beta 3) without any addons, but the problem is still there.-- Netheril96 ( talk) 03:17, 4 June 2011 (UTC)
It's good to hear that turning off Direct2D support in Firefox finally solved the problem. For any reader running into the same issue, you can do that by unchecking hardware acceleration under "Preferences" menu -> "Advanced" tab. Alternatively, there are a few switches in "about:config" dealing with "gfx" rendering issues, such as "gfx.direct2d.disabled".
For your desire of having "bolder" fonts by default, that's hardly doable because there are only two type faces, "normal" and "bold". So if you set "normal" fonts "bolder" they will become "bold" but the "bold" symbols won't become darker. The best you could do is to simulate darker characters by rendering them twice. For example, adding the following line to your vector.js will print the second run with a vertical offset of 1 pixel:
Don't think you will like the result, though. Nageh ( talk) 20:15, 7 June 2011 (UTC)
\tfrac: ['\\dfrac{#1}{#2}',1,2]
A discussion on mathJax is current underway at Wikipedia_talk:WikiProject Mathematics#MathJax.-- Salix ( talk): 07:30, 25 June 2011 (UTC)
Hello there. Firstly, thanks a lot for all of your work involving MathJax. One problem that I have notice is with subscript square bracket: . The final bracket is very faint and is almost invisible. — Fly by Night ( talk) 23:01, 25 June 2011 (UTC)
Another thing that needs attention:
Michael Hardy ( talk) 04:01, 27 June 2011 (UTC)
Another similar bug:
Michael Hardy ( talk) 23:15, 28 June 2011 (UTC)
\underline \mathbb Z_X
and \hat \boldsymbol \theta
fail in the "amsart" documentclass. As MathJax does not implement whole documentclasses there is no way of telling what would be the "right" interpretation of parameter precedence. The third problem is that \hat \boldsymbol \theta
and \hat {\boldsymbol \theta}
are typographically and semantically wrong. You do not want a bold-formatted vector that is modified with a normal-formatted hat on it, but you want a bold-formatted vector whose notation includes a (bold) hat on it. Concerning \underline \mathbb Z_X
, there may be a way to special-case \mathbb
... I'll have to investigate into this.For those using MathML to render output, I have fixed a bug concerning incorrect rendering of "special" font styles such as blackboard bold. Nageh ( talk) 18:46, 7 July 2011 (UTC)
Here's the code:
and now I put it within math tags:
The way this now appears in my browser window is that I no longer see this
but instead I see this
So the bug that was there earlier is gone, but it's still not putting the 45 points of extra vertical space between lines. There are times when that feature is needed, and it currently appears in MANY Wikipedia articles. Michael Hardy ( talk) 19:24, 30 June 2011 (UTC)
Since I started viewing things on Wikipedia via mathJax, the blackboard-bold C below appears much smaller:
Does mathJax do that somehow? Michael Hardy ( talk) 18:02, 5 July 2011 (UTC)
This one should make you happy, Michael. :) One of the MathJax developers (dpvc) has committed a patch for this issue, and I have just integrated it into my user script. Nageh ( talk) 12:58, 4 August 2011 (UTC)
In the references section to Erdős–Straus conjecture ( this version) there is a title of a paper by Tao with \frac in some math. When I view it in Chrome (on OS X) I see only " = + + "; that is, the stuff inside the fractions is missing. Maybe this has something to do with being in a small font and two column references format? Anyway it seems to be a bug. In the previous revision the math had a \scriptstyle in it but I took it out, thinking that might help; nothing changed, though. — David Eppstein ( talk) 20:30, 7 July 2011 (UTC)
I notified the mathJax developers of the existence of this present page and they seem to be looking it over. Michael Hardy ( talk) 22:55, 21 July 2011 (UTC)
Annoying as it is, I still like mathJax better than the default non-scalable, non-customizable image mode. Bugs certainly need fixing, but you should not wait until every aspect is impeccable to start using or promoting something.— Netheril96 ( talk) 05:34, 25 July 2011 (UTC)
Michael, that bug where you would actually see [10pt] has long been fixed. This was truly a show stopper, but missing the intended vertical space is more a flaw than a bug. Nageh ( talk) 08:18, 25 July 2011 (UTC)
Hi Nageh, thanks for your efforts on this. I hope this is the appropriate place to post a bug (?) report.
I followed the installation instructions. Now when I load a page with math on it mathjax tries to load, but stalls with this message:
"Loading [MathJax]/jax/output/HTML-CSS/fonts/STIX/fontdata.js"
After maybe 10 seconds mathjax gives up loading the file, gives a file-not-found sort of error, and then is stuck with the message
"Processing math: 0%"
I'm using the Vector skin, and have this in the common.js skin file:
mathJax={}; mathJax.fontDir=" http://cdn.mathjax.org/mathjax/latest/fonts"; importScript('User:Nageh/mathJax.js');
I tried it also without the first line above (starting with "mathJax") with the same result.
If I disable the STIX fonts on my computer, I do not have the problem. (I'm using a mac with OSX 10.7 (Lion) which comes with STIX fonts installed.)
Thanks Nealeyoung ( talk) 04:21, 31 July 2011 (UTC)
I've just noticed that while using mathJax,
does not look the same as
They both gave identical results before I started using mathJax. Michael Hardy ( talk) 12:57, 6 August 2011 (UTC)
IE has nasty performance issues with .offset{Width,Height,Left}, which MathJax heavily relies on. There is a ticket at Microsoft Connect that complains about this. MathJax updates the DOM immediately after each formula is typeset, typically introducing a lot of new nodes, and causes a full-page re-layout. Such quadratic behavior, exacerbated by slow accesses to .offset{Width,Height,Left}, renders MathJax painfully slow on pages like " List of trigonometric identities" which contain huge numbers of formulae.
A workaround is to prevent IE from frequent re-layouts. My approach is to let MediaWiki generate PNG as usual, do something like
$('img.tex').each(function() {
var span = document.createElement('span');
span.className = 'tex';
span.textContent = this.alt;
this.appendChild(span);
});
then register a listener for the "End Typeset" signal
MathJax.Hub.Register.StartupHook('End Typeset', function() {
$('img.tex').each(function(){
var parent = this.parentNode;
parent.insertBefore(this.children1], this);
parent.replaceChild(this.children1], this);
});
});
Since children of IMG elements do not participate in layout, MathJax's DOM updates have minimal impact on performance. This enables IE to typeset noticeably faster while allowing all formulae to be visible throughout.
Perhaps similar logic can be baked into the MathJax startup script. Kxx ( talk | contribs) 23:21, 9 August 2011 (UTC)
I recall the MathJax main developers having discussed such issues before though I haven't checked whether any relevant code has been added to the code repository. There are related discussions at the MathJax developers mailing list, e.g., [3]. You might want to bring this up at that mailing list or at the user mailing list as it is really something that should be integrated into the MathJax core code rather than my script only. Thanks, Nageh ( talk) 11:25, 10 August 2011 (UTC)
In Expectation–maximization algorithm, after the heading "Alternative description", the two lines of TeX after "expectation step" and "maximization step" are not getting rendered when I view them while logged in (and using MathJax. All I see is the TeX code. What's going on? Michael Hardy ( talk) 19:51, 5 October 2011 (UTC)
At https://bugzilla.wikimedia.org/process_bug.cgi I put in a feature request in February. This morning I looked at User:Michael Hardy/scratchwork and found that \koppa, \stigma, and \coppa within TeX are finally getting rendered! I was delighted to see this. But now it turns out that they work properly only when I not logged in! That makes me suspect the problem is with MathJax.
I need these in order to continue work on Ptolemy's table of chords. And there are probably other articles in which they should be used. Michael Hardy ( talk) 20:16, 5 October 2011 (UTC)
When I'm logged in and using MathJax, this table doesn't appear the same way it appears when I'm not logged in. I intended that between the first and second columns, and again between the fourth and fifth columns, two vertical lines would be visible. That is in fact how it appears when I'm not logged in. But when I'm logged in, and hence using MathJax, I see only a single vertical line there. The comprehensibility of the table depends on those two vertical lines being marked as different from the others.
Another difference is that when I'm logged in and using MathJax, I see vertical lines at the left and right edges of the table, and those are not there at all when I'm not logged in. Michael Hardy ( talk) 02:54, 14 October 2011 (UTC)
This looks very good when I'm not logged in, but when I'm logged in (and hence using mathJax), I get neither an error message nor the intended image; rather I just get TeX code, all on one long line, so that I have to scroll a long distance to the right to see the end of it. What's going on? Michael Hardy ( talk) 14:57, 15 October 2011 (UTC)
\overset\text{`}
, which MathJax interprets as \overset{\text}{`}
– change it to \overset{\text{`}}
, as I have done above.
Nageh (
talk)
15:30, 15 October 2011 (UTC)
Look at this:
The code between the math tags is as follows:
What appears on my browser is what I'd expect if I'd written the following:
Let's try the latter code:
This could lead to grave misunderstandings: the reader would think that what's intended is the greatest integer less than or equal to X. Michael Hardy ( talk) 14:44, 17 November 2011 (UTC)
One the N-sphere page the function \arccot is displaying in red. I guess that this is a function defined in texvc and not in MathJax. It would be possible to replace the \arccot with \operatorname{arccot} but should we be adding workarounds when texvc is still the default renderer.-- Salix ( talk): 08:29, 9 January 2012 (UTC)
MathJax crams some things together that currently has a space between. For instance in
MathJax puts no space between the second lim and x Dmcq ( talk) 17:35, 29 February 2012 (UTC)
It seems like MathJax 2.0 has indeed resolved this issue. I have just upgraded to the new release. As always, report any bugs you run into on this talk page, and thanks for testing! Nageh ( talk) 16:25, 1 March 2012 (UTC)
If an maths expression is positioned at the start of a new line because it won't fit on the last and then hover over it, e.g. try the magnified image appears at the right at the end of the last line. Dmcq ( talk) 22:50, 1 March 2012 (UTC)
I the current scheme if one makes a error in the tex the whole expresssion is shown in red in the preview. However I noticed that none of the expressions was output when I had an error and was using MathJax. For instance in
If you take the nowiki tags off the end none of the them show up whereas only the second one has an error with two back slashes. This can be a bit disconcerting needing a binary chop if the error isn't obvious. Dmcq ( talk) 11:37, 7 March 2012 (UTC)
Mathjax 2.0 changes behavior such that MJ 2.0 no longer makes use of firefox mathml by default. I am only mentioning this to you in case you want to update the firefox mathml discussion on the mathjax installation page.
http://www.mathjax.org/2012/03/02/news/mathjax-2-0-and-the-default-rendering-in-firefox/
DouglasCalvert ( talk) 03:38, 11 March 2012 (UTC)
I would like to use mathjax on the Discrete Probability page from wikibooks. Is there an easy way to do this?
http://en.wikibooks.org/wiki/High_School_Mathematics_Extensions/Discrete_Probability
DouglasCalvert ( talk) 04:14, 11 March 2012 (UTC)
Any updates on when mathJax will be ready to add to wikibooks? Creating a wikibook here:
http://en.wikibooks.org/wiki/Parallel_Spectral_Numerical_Methods
which could desperately use more LaTeX math features, in particular automatic equation numbering. Any answers would be particularly welcome in the discussion here:
http://en.wikibooks.org/wiki/Wikibooks:Reading_room/Proposals
Benson Muite ( talk) 12:39, 18 November 2012 (UTC)
Hi, Nageh. Have you seen this script from MathJax which says it takes the images, back-renders them into the tex, and calls MathJax. I'm no programmer (nor do I play one on the radio), so is this helpful to you or us? Thanks, and thanks for writing your script! -- Avi ( talk) 20:08, 12 March 2012 (UTC)
Mathjax 2.0 added a new rendering method: SVG. It looks clearer in my opinion. Any chance that you may make it functional on Wikipedia? — Preceding unsigned comment added by Netheril96 ( talk • contribs)
The following displays with the old scheme but not with MathJax
Cheers Dmcq ( talk) 14:13, 23 March 2012 (UTC)
\dot{\vec u}
).
Nageh (
talk)
15:31, 23 March 2012 (UTC)How can I use the locally-installed STIX font set? — Edokter ( talk) — 10:19, 1 April 2012 (UTC)
How can I check is mathJax loaded on a given Wikipedia page? Could you implement some "debug mode" with extra information such as "some required components loaded"… "nn chunks of math found"… "nn chunks processed"… "nn chunks contain errors"? Incnis Mrsi ( talk) 18:49, 2 April 2012 (UTC)
javascript:alert(typeof(mathJax)!="undefined"&&mathJax.loaded)
into your address bar of the current browser window to check whether MathJax has been loaded. Once MathJax has been loaded, you can right-click on any maths formula and invoke the About MathJax entry in the menu, which will give you a list of loaded configuration items.Where is (will) the base CSS for MathJax (be) located? There are some padding issues that need to be fixed, such as the excesive vertical padding around stand-alone formulas. — Edokter ( talk) — 13:10, 15 April 2012 (UTC)
<dd>
tags that they have some top and bottom margin (0.2em and 0.5em). MathJax (or rather, the .MathJax_Display class) uses top and bottom margin of 1em by default... that is very different from LaTeX.\displaystyle
.:<math>
, that is, started with a colon. What happened was the mathJax was literally interpreting the maths as to be rendered in scriptstyle font, which resulted in this tiny, illegible font. I have resolved this issue now and will update the mathJax documentation. I hope this resolves one of your issues.I'm just trying the "quick test" of your version of MathJax, since the standard (experimental) version is too obviously broken. I see that the \scripstyle
being ignored makes the integration region expression too large in the {{
oiint}} and {{
oiiint}} templates. Adding braces as stated does not appear to fix this. How to get the desired subscript size? —
Quondum
☏
20:34, 13 May 2012 (UTC)
While it looks like the experimental extension is using the Tex-AMS-MML_HTMLorMML.js configuration file, I would like to be able to use Tex-MML-AM_HTMLorMML.js so that I can also make use of asciimath2jax to display ASCIIMathML You provide the ability to tweak the Hub settings, but I think not so I can change the config file. What are your thoughts on this? Net-buoy ( talk) 23:58, 17 May 2012 (UTC)
Hello, Nageh.
There is now an experimental option in Wiki preferences to enable MathJax as the math viewer. Is this an implementation of your script, or something different, and do you recommend removing your script from JS now that there is this option? Thank you! -- Avi ( talk) 13:52, 30 May 2012 (UTC)
Whenever I edit an article with math in it, the Show preview button leaves the math portions as plain TeX in black (e.g. $ \mathbf{v} = (v_x,v_y,v_z). $). The math displays fine when I save the edit, though. Is this normal? My math appearance preference is set to Leave it as TeX (for text browsers) (although MathJax has the same problem) and my
User:Tschwenn/vector.js contains only one line: importScript('User:Nageh/mathJax.js');
--
TSchwenn (
talk)
21:59, 30 May 2012 (UTC)
javascript:alert(typeof(mathJax)!="undefined"&&mathJax.loaded)
into your address bar when you are editing an article and tell me whether it returns "true" or "false"?
Nageh (
talk)
22:23, 30 May 2012 (UTC)javascript:alert(typeof(mathJax)!="undefined"&&mathJax.loaded))
(close parentheses)? Putting either into the Firefox address bar and pressing enter does nothing at all, and in internet explorer it just redirects to the non-existent
http://www.javascript:alert..
. I tried suppressing the www.\%s.com auto-correct rule in my registry, but then IE just sent me to a Bing search for the command. On my other Windows 7 PC, both Chrome and Firefox do nothing with the command either. --
TSchwenn (
talk)
04:06, 31 May 2012 (UTC)alert(typeof(mathJax)!="undefined"&&mathJax.loaded)
and it should return either "true" or "false" (hopefully).
Nageh (
talk)
09:45, 31 May 2012 (UTC)false
everywhere, including after clicking Show preview. I have Firefox 12.0. --
TSchwenn (
talk)
12:49, 31 May 2012 (UTC)alert(MathJax.Hub.config.config)
gives, and whether mathJax.Config();
or MathJax.Hub.Startup.onload();
has any effect when you are previewing a page with unrendered math?
Nageh (
talk)
20:13, 31 May 2012 (UTC)TeX-AMS-texvc_HTML.js
. The other two commands do this in the web console: [16:30:28.821] mathJax.Config(); [16:30:28.834] undefined -- [16:30:58.644] MathJax.Hub.Startup.onload(); [16:30:58.657] undefined
. There are no apparent effects on the page. --
TSchwenn (
talk)
21:36, 31 May 2012 (UTC)MathJax.Hub.Update();
, MathJax.Hub.Reprocess();
, MathJax.Hub.Rerender();
do anything?
Nageh (
talk)
17:07, 1 June 2012 (UTC)(function anonymous() {return arguments.callee.execute.apply(arguments.callee, arguments);})
--
TSchwenn (
talk)
19:44, 1 June 2012 (UTC)Guys, the MathJax CDN serving the web fonts is currently experiencing some problem. This means that everyone who has included mathJax={}; mathJax.fontDir="
http://cdn.mathjax.org/mathjax/latest/fonts";
in his/her vector.js or common.js file will currently not be able to see any rendered math but only plain TeX code. I will inquire the MathJax folks about it.
Nageh (
talk)
18:14, 31 May 2012 (UTC)
mathJax={}; mathJax.fontDir="
https://c328740.ssl.cf1.rackcdn.com/mathjax/latest/fonts";
and it still seems to work fine. --
Avi (
talk)
18:51, 31 May 2012 (UTC)
Issues seems to be resolved. Nageh ( talk) 20:05, 31 May 2012 (UTC)
If I select the SVG option the browser outputs TeX, and I don't know how to get back to showing the formulae again with MathJax to change the option. Any suggestions? Thanks. I'm using Google Chrome with your javascript. I notice a box at the bottom saying 'Loading [MathJax]/jax/output/SVG/config.js' and then it puts out a long line with Failed to load and a full name with that name at the end. Dmcq ( talk) 23:23, 1 June 2012 (UTC)
In Squared triangular number, two of the references have mathematical formulae in them (with binomial coefficients, so impossible to rewrite as html). When I view them using Nageh/mathjax, I see a spurious italicized "< spanstyle="display:none;">< /span >" at the end of the reference. For example (with some minor changes to the TeX markup that don't affect the problem):
{{
cite journal}}
: CS1 maint: multiple names: authors list (
link)displays as something approximately like
(where I've replaced the binomial coefficient with the letter X). Oh, and it's a math italic font, not the text italic, and the angle brackets in the span don't look quite the same. Looking at the html source code generated here, I see <span class="tex"><span style="display: none;"> </span></span> which apparently is what ends up getting displayed in this way via mathjax attempting to interpret the html code as math.   is a no-break space character. Any ideas why this is happening, and how to fix and/or work around this? — David Eppstein ( talk) 21:51, 13 June 2012 (UTC)
<span class="tex"><span style="display: none;"> </span></span>
. I invite you to report this bug at
bugzilla. I am not going to do this, and other than this response I decided to be no more involved in Wikipedia activities because of the demonstrative arrogance of the WMF and its developers. It's sad, but obviously editor retention is not an issue to the WMF when it comes to established editors. HTH,
Nageh (
talk)
13:33, 14 June 2012 (UTC)Well, it doesn't happen when I link the title by hand: "Two quick combinatorial proofs of ". So maybe it is something in the cite template that is triggering this rather than in mediawiki? — David Eppstein ( talk) 16:10, 14 June 2012 (UTC)
This strongly hints to a bug in MediaWiki. There aren't supposed to be any <span class="tex"></span>
tags in the HTML source when the PNG rendering option is chosen, yet it appears in your example.
Nageh (
talk)
21:52, 14 June 2012 (UTC)
I'm finding MathJax incredibily slow (on each page load the browser takes ages loading fonts and .js files) – so slow that I set my preferences to PNG display for this reason alone. Any idea what this might be about? — Quondum ☏ 18:52, 16 June 2012 (UTC)
Would installing the STIX fonts locally help? -- Avi ( talk) 17:01, 17 June 2012 (UTC)
MathJax has this nice feature where clicking on a formula shows you a zoomed view of it. But when I click on the array in Wythoff array (in the Monobook format for Wikipedia as is my default preference), the zoomed view of the array is obscured by the Wikipedia search form, which comes up on top of the array. Is this normal? — David Eppstein ( talk) 02:18, 7 July 2012 (UTC)
<div id="p-search">
element, which prevents MathJax from overriding the layering order because it operates in a different
stacking context. One solution to solve this is to ask the MediaWiki devs to omit the z-index from the CSS... I don't see why a manual z-index would be needed at all. Another solution... well, looking at the MathJax sources it seems that the positioning element for the popup gets created as a child to the maths's immediate parent rather than as a child to the HTML root element... I guess that changing this could resolve the issue as well, but I'd need to inquire the MathJax folks on this. HTH for now.
Nageh (
talk)
10:56, 7 July 2012 (UTC)MathJax does not load in IE10 because it attempts to access document.namespaces
, which is not present in IE10. A workaround to manually create a dummy:
if (!document.namespaces) document.namespaces = { add: function() {} };
before MathJax is loaded. Kxx ( talk | contribs) 07:00, 28 August 2012 (UTC)
document.namespaces
is tested. Code that touches document.namespaces
should be removed/replaced (document.namespaces
itself is a nonstandard extension).
Kxx (
talk |
contribs)
15:46, 28 August 2012 (UTC)document.namespaces[e]
.
Kxx (
talk |
contribs)
23:00, 11 September 2012 (UTC)I switched to https mode a while ago, and the mathJax script is blocked by Chrome for it is not transferred in https. How should I modify the line importScript('User:Nageh/mathJax.js') so that the script is transferred from https://en.wikipedia.org/User:Nageh/mathJax.js instead?- Netheril96 ( talk) 14:36, 31 August 2012 (UTC)
I'm noticing a bug in MathJax, whereby the a is bolded italic, and the b is unbolded italic in <math>\mathbf{ab}</math>
:
In PNG it is it is the same as for <math>\mathbf{ab}</math>
(bolded roman), here also for MathJax:
Any ideas? — Quondum 06:34, 10 September 2012 (UTC)
<math>\mathbf{ab}</math>
is bold italic in MathJax, and bold roman in PNG, which is of course still a problem. I'm not sure whether \bold
is intended to be italic, but they should not be different. —
Quondum
13:59, 10 September 2012 (UTC)While I'm logged in, and looking at socle (mathematics), I see this line and it looks fine:
But looking at the same page when I'm not logged in, I see nothing but error messages where that line should be. What's going on? Michael Hardy ( talk) 16:39, 21 November 2012 (UTC)
I reported this at Wikipedia_talk:WikiProject_Mathematics#Math_Jax_problem
I switched over to Math Jax. On my browser (Windows IE 9, version 9.0.8112.16412, update 9.0.13) the formula
does not render correctly (the \not\equiv should be ≡ with a line through it, but is just a space on my browser)
Thanks.
And thanks for all the work on Math Jax; it looks much better than what I was getting before!
- Virginia-American ( talk) 17:04, 3 March 2013 (UTC)
Is this intended behaviour? It's not uncommon to see the markup, but then it disappears, all I see is horizontal lines where the fractions should be when the MathJax status goes away. Reloading doesn't seem to fix it, but purging does - implying it's server-side. Not a big deal for someone who's heard of purging - not so good for those who haven't. ChrisHodgesUK ( talk) 12:20, 24 June 2013 (UTC)
The page uses the typewriter font, but apparently [MathJax]/jax/output/HTML-CSS/fonts/TeX/Typewriter/Regular/Main.js is missing. Kxx ( talk | contribs) 07:47, 21 August 2013 (UTC)
User:Nageh/mathJax.js contains mathJax.version = "0.2.3";
, while
User:Nageh/mathJax/MathJax.js contains if(!MathJax.Hub){MathJax.version="2.2";
...}
. The latter is what is displayed when I right-click on some math and choose "About MathJax," so I assume it's correct. Why do these version numbers not agree, and does it matter for correct operation? --
SoledadKabocha (
talk)
05:22, 15 February 2014 (UTC)
Hi, it's not obvious to me what, if any, relationship there is between your user script and the experimental MathJax rendering option in Preferences, and the mw:Extension:MathJax. Does Wikipedia use that extension? I'm guessing no, but wanted to make sure. Also, where does discussion of math rendering on Wikipedia take place? I am wondering if MathJax will ever be turned on by default, and have no idea where to ask the question. Thanks! Klortho ( talk) 02:45, 24 March 2014 (UTC)