# List of recurrent technical problems involving MathJax

Edit: I have retagged this question 'feature-request' in an attempt to expedite action on these issues. I request that the MathOverflow developers implement an automatic fix of the issues in the answers below, together with the problem involving \*, and the problem described in this question linked immediately below.

I have previously made a request for an automatic fix to an outstanding MathJax issue involving backticks. That request has not been addressed yet...

In any case, I have come across some other issues which cause the MathJax to not render or to render incorrectly. One of them is the appearance within MathJax code of \* instead of \ast (as quid indicates in a comment below, using \ast is probably a safer option than *). I have fixed a few instances of this problem myself and today (August 7, 2013) there have been at least two questions edited to fix this same issue.

In view of the above problems involving MathJax, I am asking everyone to list other similar recurrent mathjax problems. I am looking for problems which verify all (or at least most) of the following conditions:

• It should be a technical problem involving MathJax; this is probably not the right place to deal with misuse of MathJax.

• The problem should occur in several instances.

• The problem would ideally admit a fairly straightforward automatic fix, e.g. by applying a search and replace algorithm to the text in questions and answers.

Just to be clear, the MathJax problems need not have been caused by the site update.

The ultimate goal is to have some automatic fix applied to each problem listed in the answers in this thread. This would probably depend on the intervention of the MathOverflow developers. Even if that does not happen, it may still be useful to gather a list of the problems to look for.

• It might be slightly safer to use \ast instead of *, the latter having a markdown-meaning that could interfere at some point (even if it might not at the moment).
– user9072
Aug 8 '13 at 15:44
• I've pinged the SE folks again regarding this thread. Sep 12 '13 at 23:01
• We're looking into this - sorry for the delay in our response. Sep 13 '13 at 16:56
• Dear @Tim Post: Thank your for the update. Please keep us posted of any developments. Sep 13 '13 at 17:30
• Dear @Tim Post, any updates? Nov 18 '13 at 23:17
• @RicardoAndrade I'm terribly sorry, it's been one of those weeks, I'll follow up on this first thing in the morning. Thanks for reminding me! :) Nov 19 '13 at 0:52
• Dear @Tim Post, any more updates? I apologize for the constant nagging. Jan 10 '14 at 23:55
• @Scott Morrison, sorry to bother you. Do you happen to know whether this is moving forward or not? Thank you very much. Feb 2 '14 at 19:00
• @RicardoAndrade, not to my knowledge. Feb 3 '14 at 0:55
• @TimPost Ping. The issue is still unresolved after 2 1/2 years. Jan 28 '16 at 11:36

A couple I have noticed:

• using \\, instead of \, splits the line,
• using \\; instead of \; splits the line, and
• using \\! instead of \! splits the line.

Added later: There have been some instances where people have used \_ instead of _ but I believe all of these have now been fixed. Edit by R. Andrade: Apparently, there are still instances of this issue left even though none show up using the site search function. For example, see George Lowther's comment below.

Added later: As quid points out, there was also an issue where people had used \\# instead of \# but all of these cases have been fixed. Edit 2 by R. Andrade: It appears that this issue is not resolved either. Ironically, I have just fixed (August 19, 2013) this same problem in one of my own answers.

Added later: Some people have used $$ and $$ instead of ( and ). These do not render properly.

• Another one of this type, I actually noticed I think via one of your recent edits, \\# instead of \# to get # (breaks MathJax rendering). I think these are all transition problems because before one needed the extra backslash while now this is not needed anymore and its presence causes problems. (Somewhere there was a thread on meta already on this double-backslash problem.)
– user9072
Aug 8 '13 at 15:42
• The other thread is meta.mathoverflow.net/q/238 . Aug 8 '13 at 16:13
• Dear @Michael: Thank you very much for your contribution. Aug 8 '13 at 17:55
• It appears there may still be some instances left of \_ being used instead of _. For example, George Lowther fixed such problems in his answer mathoverflow.net/a/54531/21095 one hour ago. Aug 13 '13 at 3:32
• @Michael: What method do you use to find questions with a given issue and determine if there are any left? Do you simply use the search function on mathoverflow? Aug 13 '13 at 3:44
• Yes. The appropriate search only shows one post using \_. In this case the user intended to get the resulting output. Aug 13 '13 at 3:45
• @Michael: Does that mean that George Lowther's answer did not appear in a search for \_ before it was corrected? Aug 13 '13 at 4:17
• It did not. I'm not sure why. Aug 13 '13 at 5:17
• @MichaelAlbanese Please stop these trivial edits. You edited an answer on mathoverflow.net/questions/25374/duplicate-detection-problem which had zero upvotes (while other answers had many more, including an accepted answer), thus bumping a three-year old long-settled question to the front page. There is no reason to do that. Aug 13 '13 at 17:26
• btw, I still have answers containing \\, instead of \, and \_ instead of _ (also  \\\\  instead of  \\ , but I'm not sure if they are such an issue). I just edited one, and was about to edit another but held off as I don't want to bump lots of old posts up. Aug 13 '13 at 20:58
• @Felipe Voloch: While the bumping to the front page is inconvenient, please consider that these answers and questions do not render correctly. /// @George Lowther: Thank you very much for that information. Interestingly, searching for \\, returns none of your posts, thus showing that the search function is not entirely effective in locating these problems... Aug 17 '13 at 5:42
• @RicardoAndrade All am I asking is to use some common sense and ask whether that particular question/answer really needs that edit. My example is a clear example that doesn't. Now, if the whole page is not rendering because of one faulty equation at the very bottom, I think this is a software bug and should be addressed as such. Aug 17 '13 at 14:57
• This issue still seems to exist in some places; see Doubled backslashes in old posts. Mar 8 '18 at 11:22

Since people are randomly listing various combinations of \\ with other characters, let me for the record repeat here what I wrote in the comments to Double backslashes:

• Every \\ needs to be simplified to \ unless it appears in a \substack or inside a matrix or array-like LaTeX environment (e.g., matrix, align, gather, cases), where it could be legitimately used as a line break. What follows the \\ is otherwise irrelevant, though as a particular case, \\\ or \\\\ should yield \\.
• For completeness, I want to remark that a double backslash \\ creates a new line when mathjax is rendered using CSS or SVG. Interestingly, it appears to be ignored and not create a new line when mathjax is rendered using mathml. Sep 6 '13 at 18:34

Some cases which seem not to have been dealt with so far:

$and$


for generating square brackets (as in the OP Inverting the Weyl Character Formula )

and

i &lt; j


(as in https://mathoverflow.net/revisions/82322/1 ). About the latter syntax, I'm surprised that it has ever worked (in LaTeX, not html!), but if it did, it's probably worth searching for it. As for the former one, I would suspect it was used often enough. I'm not sure if backslash-squarebracket actually is a meaningful command in LaTeX for generating square brackets, but people have definitely used it (maybe to get past markup?).

• In actual LaTeX, $...$ is a glorified version of $$...$$ (which however appears not to be supported by MathJax, at least in the local setup). In SE 1.0, the backslash would be eaten up by Markdown before being passed to LaTeX as in the other cases. Sep 2 '13 at 17:43
• Just to be sure, it may be a good idea to further search for other html symbols appearing within mathjax code. If &lt; was used for some reason, other html symbols may have similarly been used. Sep 7 '13 at 12:41
• I can confirm that html entities other that &lt; appear a few times within mathjax code in old posts (probably from before mathjax was enabled on MathOverflow). Hopefully, any automatic fix which is eventually applied to the issues in this thread will also search and replace these symbols. Dec 2 '13 at 15:43