Strange. They’re identical for me.

]]>Following David’s comment about Chrome I tried it on my machine and the letters do look similar but the J has a loop on the bottom whilst the I does not. The two letters are not looking the same.

]]>Here is the earlier thread. As in #6 there, there is a significant difference between the way MathJax is being used on MathOverflow and here; MathJax is parsing LaTeX directly on MathOverflow, whilst it is parsing MathML on the nLab.

Looking at the HTML source, it looks like the MathML produced has `<mi>I</mi>`

, where the I is already a unicode calligraphic symbol. It might be that if one used `<mi mathvariant="script">I</mi>`

or something instead, using only an ASCII I, then MathJax would render it correctly. This would involve changing the Itex2MML source, though. For now, I would suggest just to change the page so that the two letters are distinguishable, e.g. by not using mathcal or by using different letters.

I am using Chrome too, this is probably the problem. If I remember correctly Chrome stopped supporting mathml a while back, so is this really a MathJax problem? In which case Richard could reconfigure it? I am only guessing because MathOverflow uses MathJax aswell if I remember correctly.

]]>I’m using Chrome. At MathOverflow they differ for me.

]]>FWIW: On my MacBook using Firefox, they display correctly both on the nLab page and here. I tested on Safari and although looking different from here, they are distinct, and readable, in both the Lab and here.

]]>They look different to me, and each is the same in both places, neither being a unicode box.

]]>This is to do with the mathcal font being weird. I’ve brought this up before. For some reason mathcal on nlab is this weird scripty version. It should really be like it is on mathoverflow.

]]>and here?

]]>Do other people see these final symbols as identical at Serre intersection formula?

]]>Given a regular scheme $X$ and subschemes $Y,Z$ with defining ideal sheaves $\mathcal{I},\mathcal{J}$…