# Issue with vertical positioning of text with numbers

It seems there is something wrong with the text positioning on MSE site (seems alright on other sites as far as I've tried, including mathoverflow.net). Specifically if you type the following text in post/comment, you can see the issue:

Some other text around 3366 and more text.

It does not happen for numbers typeset in math mode (). I can see this result in Chrome and Microsoft Edge on the desktop. It looks alright in Chrome when viewing on the phone.

By the way the original scenario when I encountered this was referrence to OEIS sequence A333666. Looking at how OEIS sequences names are typeset on the site, this seems like a corner case where people use both styles (text and MathJax).

• Are you getting at the fact that bits of the 3 sit below the line, like a g, while the 6 sits above parts of the text, like a b? If so, this is deliberate, and based on the $\TeX$ idea that "maths numbers" are different from "text numbers", so are typeset differently. You can read about some of the logic behind this here. Sep 10 '20 at 10:41
• @user1729 But these are "text numbers", not "math numbers". It might be deliberate, but then I guess it might be caused by a font, since on other sites which are also heavy on math content this does not occurr, as well as phone vs desktop (so this is incosistent at least). By the way it looks like $33^{66}$, if that is deliberate... ok "great" :).
– Sil
Sep 10 '20 at 10:50
• Okay, I've dug around a bit more. The basic reason is: Math.SE uses the Georgia font, and this is how numbers are displayed in this font. Figures which sit like this are called text figures. It seems like this isn't standard in $\LaTeX$, so probably part of my comment above is not correct. However, one of the reasons I like the figures being like this is because I like the quote from Knuth (in the above link) about the difference between "maths numbers" and "text numbers". Making the text numbers look like text is, I believe, a good thing. Sep 10 '20 at 11:07
• @user1729 Thank you for that information, that explains where it comes from. Although that is subjective, I respectfuly disagree this is a good thing, at least in this site context (we have MathJax for math numbers anyway). I would argue this might get confusing at times, especially because how it looks like exponents, but that is a problem only if math mode is not used (so mostly not a big deal), so I don't know. I guess If community is fine with it, it's okay.
– Sil
Sep 10 '20 at 11:16
• Sep 11 '20 at 9:53

The typeface used in posts at Math.SE is Georgia. Maths is done via MathJax, and here the specific typeface used is complicated (it seems like it can change from machine to machine, but really I don't know).

Georgia uses text figures, which are designed to blend into continuous text. Therefore, the "issue" you are seeing is really a "feature". In contrast, MathJax figures are designed to stand out from continuous text.

In the context of our site, this makes sense as (when used correctly) it allows us to distinguish "text figures" from "maths figures". For example, it makes the following silly sentence easier to read:

I came across the following problem from 2009: $$2009\times 7=?$$

By "used correctly", I mean "use MathJax if and only if you are writing maths". I gave more details on the differences between "maths numbers" and "text numbers" here (which includes the interesting quote from Donald Knuth, who wrote $$\TeX$$, I mentioned in the comments to the question), and you can also read this Tex.SE thread.

I should end by noting that it seems that the default typefaces in LaTeX do not distinguish between texts figures and maths figures (and also by saying that I am not an expert in typesetting, just an interested observer, so please correct me if you spot any errors or misunderstandings on my part).

• Right, so the real question should be about the font and possible switch, and by the question you linked in comments, that discussion has been done some long time ago... Also the mathoverflow seems to use Arial, and phone version of MSE probably too (not sure why, but it is not the only incosistency, the quotes are also rendered differently).
– Sil
Sep 11 '20 at 10:58