Since our math editor is currently on the news, here's one aspect worth considering.

One thing that's been pointed out on the Meta.MathOverflow threads (here and here) is that v3 of MathJax has been released recently (IIUC last April) and that it is faster than the current v2.7. (MathJax v3 is also discussed on Meta SE in this thread, which hasn't received significant developer attention.) It doesn't seem to break anything we rely on, but looking at the list of changes, one thing that sticks out is the addition of the braket extension, which could potentially be very useful here.

The last time we discussed something similar was the physics extension, and there it does look like there are good reasons not to include it, but braket is smaller and nobler and it could do quite a bit to help with on-boarding.

In any case:

  • Assuming a switch to v3 is coming, would the braket extension make sense?
  • Would it break any existing posts which define their own bra/ket notation?
  • Would we want to have it enabled?
  • 3
    $\begingroup$ There are a few hundred posts that define their own bra or ket: data.stackexchange.com/physics/query/1369090. I'm not sure if they would break or not from adding this extension, but fixing them should be doable in any case. $\endgroup$
    – Chris Mod
    Commented Feb 4, 2021 at 5:48
  • $\begingroup$ @Chris: They shouldn't break. MathJax allows the redefinition of existing commands (even using \newcommand; AFAIK it treats \newcommand and \renewcommand identically), so the custom definitions will just override the extension. $\endgroup$ Commented Feb 8, 2021 at 12:08
  • 3
    $\begingroup$ @IlmariKaronen I think you’re right, but “shouldn’t break” is wishful thinking unless there is testing involved. $\endgroup$
    – rob Mod
    Commented Feb 10, 2021 at 23:03
  • 1
    $\begingroup$ What's the difference between using that and "|\psi\rangle" ($|\psi\rangle$)? Does it just make it a tad quicker to write that? If that's all, then is it worth it if it there's even a chance of breaking all those posts noted above? $\endgroup$ Commented Feb 12, 2021 at 0:50
  • $\begingroup$ @JohnDumancic Even if every single one of them breaks (and it's likely they won't), I don't consider that to be a high cost to pay (fixing <300 posts will only take a few days if people pitch in). If there are no concerns about the stability of the braket extension, then the only concern I see is compatibility with other SE sites. (But, as noted before by others, perhaps that should not be a goal.) $\endgroup$
    – prolyx
    Commented Feb 15, 2021 at 3:51
  • 1
    $\begingroup$ @JohnDumancic indeed, with single bras and kets there's not much difference. But some complicated brakets, especially those containing fractions (or differentiation operators) take some careful playing with \left and \right for all components (two angles and two mids) or resorting to \big etc.. Just using \Braket should be much easier. $\endgroup$
    – Ruslan
    Commented Feb 15, 2021 at 14:35

1 Answer 1


Just as a place to gather votes as a clear signal:

Yes, the braket extension is appropriate and useful. It should be enabled on this site whenever the MathJax version is upgraded to v3.

(Vote up if you agree, down if you don't.)


You must log in to answer this question.

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