Background [TL;DR]

In my previous meta question Vertical space between superscript and square root sign, I attempted (not so successfully) to illustrate a display problem about the square root sign by $\sqrt{e^\overline{X}}$: $\sqrt{e^\overline{X}}$.

I had thought about enlarging this: ${\Huge \sqrt{e^\overline{X}}}$: ${\Huge \sqrt{e^\overline{X}}}$, but the stroke going northeast $\nearrow$ surpasses the horizontal stroke going to the right-hand side $\to$.

screenshot on opera
Screenshot using Opera 71 on Ubuntu 20.04 LTS

MWE (min. working example)


$$\sqrt{a}$$ $$\Large{\sqrt{a}}$$ $$\Huge{\sqrt{a}}$$


square root display problem by Firefox
Screenshot taken by Firefox 84.0.1 on Ubuntu 20.04 LTS

Expected output

The right-upward directed stroke $\nearrow$ and the horizontal stroke $\to$ are perfectly joined together as shown in the following image, but without the vertical stroke $\downarrow$.

expected output
image source: Badly scaling closing SQRT symbol with large font sizes on $\rm\TeX$.SE

  • 3
    $\begingroup$ I see the same on Chromium 87.0.4280.88 (Official Build) Arch Linux (64-bit). $\endgroup$ – Asaf Karagila Jan 8 at 13:16
  • $\begingroup$ @AsafKaragila Thanks for comment. I've edited my question to include the expected output, which I had omitted carelessly. $\endgroup$ – GNUSupporter 8964民主女神 地下教會 Jan 8 at 13:34
  • 1
    $\begingroup$ possibly related - github.com/mathjax/MathJax/issues/1044 $\endgroup$ – Anindya Prithvi Jan 9 at 16:11
  • 1
    $\begingroup$ Totally off-topic. In order I "see" and enjoy the "name" my wife did print it using a 96 font size, framed it for decorating my office. $\endgroup$ – Claude Leibovici Jan 14 at 15:29
  • $\begingroup$ Maybe something can be achieved using \strut or \mathstrut, but the closest I got is \Large{\sqrt{\Huge{\mathstrut a}}} $$\Large{\sqrt{\Huge{\mathstrut a}}}$$ which is a bit emaciated. $\endgroup$ – PM 2Ring Jan 18 at 2:42
  • 2
    $\begingroup$ It turns out that this is a bug in the HTML-CSS output of MathJax, where one term in the positioning of the overline doesn't take the scaling into account, so the line is misplaced, and the error is more pronounced the larger the scaling factor is. I have made a issue and accompanying pull request to resolve the issue. It will be fixed in the next v2 release (though StackExchange is using v2.7.5 rather than the current v2.7.9). $\endgroup$ – Davide Cervone Jan 20 at 13:31
  • $\begingroup$ I wonder if there's an OS / font / screen resolution dependency too? I'm on Firefox 84.0.2 on MacOS Big Sur (and using the HTML-CSS math renderer) and all the examples on this page look fine to me. I do see a slight discrepancy in the overline height on Chrome, but not nearly as much as in the screenshots others have posted. Weird… $\endgroup$ – Ilmari Karonen Jan 21 at 3:26

Neither a solution nor an analysis of the inner working of MathJax. Just posting


for comparison:


This looks better on my screen (Firefox, version 84), but may be is not quite what you wanted.

  • $\begingroup$ Thanks for response. I go with everything in this answer expect the last half-sentence. $\endgroup$ – GNUSupporter 8964民主女神 地下教會 Jan 10 at 15:15
  • 2
    $\begingroup$ See Davide Cervone's reply to a related question. In particular for the description of the surd sizes. He works for MathJax, so knows what's going on. The reasoning that lead me to try this is the following. The size of the surd \sqrt tries to automagically match the box that follows. So if Huge is desired, we only need to make the contents of that box huge, and the surd will follow suit. In light of that it does not really make sense to try and apply a modifier to the surd itself. $\endgroup$ – Jyrki Lahtonen Jan 15 at 7:46

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .