# Do we need LaTeX formatting?

Is LaTeX formatting desirable in German Language SE? Just tried to answer a question which involved a formula, but when I answered, it did not work. In the question, though, it comes up correctly.

No, it does more harm than good.

The question you are referring to is probably the only one on this site where MathJax would have been a slight benefit (and I fail to imagine a question for which it would be actually needed). Everywhere else, it would be disadvantageous: It increases the loading time and it may be activated accidentally and confuse users. There are many many sites on the network, which would need LaTeX more than we do, and do not have it (probably for good reasons), e.g., StackOverflow and TeX.SE (yes, you read that right).

Since this question got a late answer saying it might help, I am going to post a second late answer in line with what @Wrzlprmft posted more or less.

Bottom line: We could include it, but it would be of exceedingly small value and therefore we might as well leave it.

Reasoning:

• We are a language site. The vast majority of our questions revolved, is revolving and will revolve around pronunciation, grammar, style and translation issues. The number of questions actively asking about anything that needs mathematical symbols will always remain low whether we are talking about one, two or ten.

• The first question linked could have been formatted entirely without mathematical symbols by using mere placeholders.
The second question linked explicitly asked how to read (i.e. pronunce) certain symbols. Any further questions asking the same thing should be forwarded to that one as duplicates.

• Even experts who regularly use TeX syntax on other Stack Exchange sites (e.g. me and Loong) will probably struggle to remember situations on here where they actually wanted to use TeX. (Not counting the two linked questions.)

• The most complicated symbols we will use are probably going to be IPA symbols. However, to the best of my knowledge there is no easy way to input IPA symbols via the MathJax extension open to Stack Exchange sites. So it won’t even be any easier to input formulae.

• The Unicode workaround is not reasonable for users not knowing about the workaround since it binds workpower in regular users who do the substitutions ex post.

In this context, I highly disagree with @hier’s argument for two reasons:

1. There won’t be many cases anyway, so those three that actually arrive can be error-handled by us.

2. New users still need to get shown how to do it properly. From my experience over at chemistry, ‘new users’ — i.e. those with bad MathJax formatting — includes users up to 1000 rep.

We also have a significant number of regulars who go through almost every post correcting apostrophes, dashes and quotation marks to the typographic variants. I’m sure one of those will volunteer to take the burdon of adding the unicode-workaround symbols without being asked.

• Remember that the extension we are talking about is MathJax, not LaTeX! It is designed specifically for mathematical environments and includes automatic maths formatting — e.g. italic letters. Not every LaTeX package is or can be supported by MathJax. Specifically (examples taken from chemistry.SE), mhchem works while upgreek does not. Chemistry, of course, has a dire need for mhchem support to typeset chemical formulae — although that can also be done nicely via HTML with some work: see H2SO4 — but would also love upright Greek letters for a variety of purposes. The latter either have to be entered as is (i.e. as their unicode codepoints) or via a rather complex syntax or end up italicised (which would be wrong by nomenclature standards). As soon as the domain of maths is left, the MathJax extension becomes exponentially less helpful.

• And just to clearly state a consequence of the previous point: Enabling MathJax would not allow for any easier formatting of umlauts or ß for users without German/Austrian (/Swiss for umlauts) keyboards. If they were to use $\"a$, that would not give ä!

In consequence that means enabling MathJax would be of little to no benefit. Since we are Germans and thus cherish the idea of benefit-cost analyses, we must arrive at the conclusion that no benefit means no action need be taken.

• Alright. You convinced me. At least y'all oughta be proud o'me that I didn't start a dup but searched for this thread first! :D – hiergiltdiestfu Oct 22 '15 at 20:37
• I just realised I was wrong. $\"a$ would not give *ä* at all since we are in a maths environment. Instead, it does not recognise \" as a proper macro, prints \” in red and a following it in black. (Just tested on Chem.SE.) – Jan Oct 24 '15 at 22:55
• Well, that basically invalidates your entire argument :-) – hiergiltdiestfu Oct 25 '15 at 16:06

Yes, it is a good idea.

• It helps with more than just questions regarding pure mathematics because it makes it easy to insert certain symbols and formulae in general. This is applicable for all fields using special compound symbols, like for example Computer Science and Engineering.

• The number of questions relying on TeX syntax is rising. "the only one on this site" is not true anymore. Recent example.

• It would be much more welcoming to academics and engineers who use TeX on a day to day basis and can expect that to work on every SE site if they know it works on certain SE sites. (I know I was heavily confused that it worked on some but not on all.)

• Granted, Unicode is a workaround for single symbols, but I postulate that the rendering result of the TeX support library is more widely displayable than the full set of Unicode.

• The Unicode workaround is not feasible for formulae or questions including lots of unique TeX.

• The Unicode workaround is not reasonable for users not knowing about the workaround since it binds workpower in regular users who do the substitutions ex post.

• The data from the counter-argument regarding performance, which is linked in Wrlz's answer, is two years old. Chances are the inherent performance of the rendering lib, the performance of JS engines in browsers, and the general performance of the host machines have all improved exponentially ever since.

• There's no reason to restrict technical possibilities just because it could confuse users, since users can be expected to expect certain formatting syntax to be unknown to them at first, as with every other forum software. Chances are, the rise and prevalence of Blogs and Markdown-based sites have improved the expectations of potential users in that regard over the years already.

At least, the performance profiling should be repeated and professionalized such that the performance impact can be reassessed.

If a site-wide solution is not feasible, an optional per-question inclusion of that dependency should be explored, as proposed in this answer.