tea.mathoverflow.net - Discussion Feed (eucal and mathrsfs in jSMath?)

Anton Geraschenko comments on "eucal and mathrsfs in jSMath?" (4255)

Anton Geraschenko — Tue, 30 Mar 2010 15:51:50 -0700

Oh, I see. Yes, that's probably possible. You'd have to do jsMath.Font.Load("eusm10"); and (I think) jsMath.Parser = {prototype:{macros: {mathcal: ['Macro','{\\eusm #1}',1]}}}; but you'd need to do this before jsMath.Process(document) is executed.

Andrew Stacey comments on "eucal and mathrsfs in jSMath?" (4252)

Andrew Stacey — Tue, 30 Mar 2010 13:57:18 -0700

What I meant was that fpqc should be able to redefine stuff so that firefox serves up eucal instead of CMcal just for him without disturbing the rest of us.

"Client-side" means "Anton doesn't do anything".

Anton Geraschenko comments on "eucal and mathrsfs in jSMath?" (4225)

Anton Geraschenko — Mon, 29 Mar 2010 18:07:26 -0700

I can do it client side by redefining the macros, but the point is that most people who have installed extra fonts have installed CMcal, but not eucal, so they'd have do download the image fonts for eucal.

Andrew Stacey comments on "eucal and mathrsfs in jSMath?" (4214)

Andrew Stacey — Mon, 29 Mar 2010 14:31:47 -0700

You should be able to do this client-side. I've no real idea how to do it, but if you've downloaded the fonts you ought to be able to muck around with them to your heart's content.

Harry Gindi comments on "eucal and mathrsfs in jSMath?" (4143)

Harry Gindi — Sat, 27 Mar 2010 14:09:38 -0700

Nothing in particular, I just like eucal better =). It seems like it's a lot of trouble to go to, and since you're not going to enable the other fonts in general, it doesn't seem like it's worth it. Nevermind!

Anton Geraschenko comments on "eucal and mathrsfs in jSMath?" (4141)

Anton Geraschenko — Sat, 27 Mar 2010 13:50:15 -0700

Not cleanly. cmsy10 is loaded by default (as are a bunch of cm fonts). I could load eusm10 and redefine \mathcal and \cal, which is pretty crufty. It would also be annoying for people who actually went to the trouble to install the TeX-fonts to get faster pages or cleaner type, because cmsy10 (which contains CMcal) is included in the TeX fonts but eusm10 is not.

What's wrong with CMcal?

Harry Gindi comments on "eucal and mathrsfs in jSMath?" (4138)

Harry Gindi — Sat, 27 Mar 2010 12:39:39 -0700

Could you swap CMcal for euler script?

Anton Geraschenko comments on "eucal and mathrsfs in jSMath?" (4136)

Anton Geraschenko — Sat, 27 Mar 2010 12:15:57 -0700

We should aspire to make the site as fast as possible for everyone.

Even if speed were not an issue, it would pain me to see somebody asking a question using eucal and somebody else responding with mathcal. I still can't come up with any good reason to support multiple calligraphic fonts.

Harry Gindi comments on "eucal and mathrsfs in jSMath?" (4130)

Harry Gindi — Sat, 27 Mar 2010 08:25:20 -0700

Anyone with that slow a connection is going to have to wait a looooooooooooonnnnnnnnnnnng time to do anything, so he/she will be used to it.

There's also an option to turn jsMath off totally, in which case you can just read the LaTeX.

Andrew Stacey comments on "eucal and mathrsfs in jSMath?" (4129)

Andrew Stacey — Sat, 27 Mar 2010 07:32:27 -0700

You can download the opentype or truetype fonts directly to your computer, in which case, it is not an issue!

Yes, I know that I can do that. But only die-hards like those that frequent these metaphoric meeting halls are going to both know that they can do that and be bothered to do it. I'm concerned about those who encounter MO for the first time, possibly over a slow connection, and find that it takes a loooooonnnnnnggggggg time to see anything.

Harry Gindi comments on "eucal and mathrsfs in jSMath?" (4128)

Harry Gindi — Sat, 27 Mar 2010 06:27:35 -0700

You can download the opentype or truetype fonts directly to your computer, in which case, it is not an issue!

Andrew Stacey comments on "eucal and mathrsfs in jSMath?" (4127)

Andrew Stacey — Sat, 27 Mar 2010 06:13:24 -0700

I don't believe either of those statements either. (There's a sentence with two "either"s which hopefully mitigates the appalling sentence I just posted containing two "actually"s.)

I would like to point out a danger of having two calligraphic fonts available. That is, that while it is unlikely that one person will use both, there is usually more than one person involved in what's on a page. So all it takes is for one person to use one font and another to use the other and you've got both being downloaded.

(Incidentally, when the fonts download, do they download the whole font set or just the characters being used?)

Anton Geraschenko comments on "eucal and mathrsfs in jSMath?" (4124)

Anton Geraschenko — Fri, 26 Mar 2010 23:30:15 -0700

Well, I guess I can't blame Scott for being out of town any more. I'm still not sure what to do. I need somebody to convince me that either

we need to have more than one calligraphic font, or
loading an additional font slows things down even if a page doesn't use that font.

I don't really believe either one of them right now. For the first one, I think it would be bad form to use multiple calligraphic fonts in the same post (after all, if it turns out to cause speed issues, we'll probably make them all synonymous, making such a post very difficult to understand), and I don't see why somebody would really want to use one and not another. For the second, I'd like to have a test I can perform that gives me reproducible numbers (or at least a reproducible difference or ratio).

Scott Morrison comments on "eucal and mathrsfs in jSMath?" (4118)

Scott Morrison — Fri, 26 Mar 2010 21:22:49 -0700

@Anton,

it appears that we already had rsfs10 in latex.mathoverflow.net/jsMath/fonts/, but not eusm10. I've fixed that now. I haven't done anything about macros. I think you should now have sufficient privileges in the latex.mathoverflow.net/ hierarchy.

Andrew Stacey comments on "eucal and mathrsfs in jSMath?" (4102)

Andrew Stacey — Fri, 26 Mar 2010 16:49:45 -0700

Just tested it. On my recent question, http://mathoverflow.net/questions/19475/can-i-detect-the-point-of-impact-without-looking-at-it, over a reasonably slow wireless in Copacabana (anyone jealous?), it took 20s to finish loading the page (no answers yet). That is without having to download any fonts, since I've already looked at that page several times so everything that could be cached ought to already be so.

Andrew Stacey comments on "eucal and mathrsfs in jSMath?" (4101)

Andrew Stacey — Fri, 26 Mar 2010 16:40:20 -0700

I suppose I ought to download those fonts (darn it, I've already got the STIX fonts, what more does jsMath want from me?). However, I'll need to figure out if it can be done in such a way that clearing my cache doesn't get rid of them (anyone know? And, no, I'm not going to turn off that feature.).

However, you can be pretty sure that I'm going to stick around MO for a while even if jsMath proves such a pain. But it's not me that you should be worrying about: it's the ones who come for the first time, find that it takes half an hour (yeah, yeah) to load even the first page, and go away again.

Harry Gindi comments on "eucal and mathrsfs in jSMath?" (4056)

Harry Gindi — Wed, 24 Mar 2010 19:12:32 -0700

@Andrew: If you download the fonts for jsMath, it does not have to load image fonts, so you're golden!

Andrew Stacey comments on "eucal and mathrsfs in jSMath?" (4055)

Andrew Stacey — Wed, 24 Mar 2010 19:02:31 -0700

Kevin, the abbreviations aren't the problem, it's the extra junk that has to be sent over the wire in order to display it.

(Really, all the conversion should be done server-side but that's not possible with the current software so I'm not going to shout about that. I only mention it now because I've been hacking together some MathML-generation scripts and it's on my mind.)

(Oh, and I loathe tex files that start like that. At the very least, they ought to use a loop command to define all the commands in one go.)

Anton, I'll see what I can do when I get back to Norway. I'm currently in Brazil on an extremely buggy and slow connection via a hotel wireless network (at least, I hope it's the hotel's network. If next time I answer a question then I appear to be in the Antarctic then you'll know that it wasn't the hotel's network and someone's stolen my id) so these issues are even more important!

Being away from my desk and trying to do mathematics over slow connections and via non-university subscribers has reminded me that places like MO have extra value because they are completely open. The only bar to joining in is understanding the mathematics: there's no subscription or limitation on who can ask or answer. Let's not get rid of that by putting in loads of stuff to download which is fine for those in the US with super-duper connections, but makes it next to impossible to use for those with more shaky infrastructure.

Mariano comments on "eucal and mathrsfs in jSMath?" (4053)

Mariano — Wed, 24 Mar 2010 16:05:35 -0700

The problem is that we all have different abbreviations!

Kevin Buzzard comments on "eucal and mathrsfs in jSMath?" (4052)

Kevin Buzzard — Wed, 24 Mar 2010 15:22:19 -0700

If I'm answering a question on MO it's typically a number theory question and I find myself having to write "\mathbf{Q}" 6-7 times. This actually put me off once to the extent that I just typed a plain text answer, writing Q_p instead of $\mathbf{Q}_p$ on the basis that anyone reading would know exactly what Q_p meant and who cares if it didn't look as good---indeed, if some did care they could just edit the post themselves :-)

My point then is that another solution would be to define some standard macros: \bfQ for \mathbf{Q}, \calC for \mathcal{C} etc etc, that all users could use. Would this be a solution to fpqc's underlying problem which would make Andrew happier? I have seen TeX files by other people which start with "\def\bfA{\mathbf{A}}\def\bfB{\mathbf{B}}...\def\bfZ{\mathbf{Z}}\def\calA{\mathcal{A}}..."even if most of the defs aren't used in the paper.

Anton Geraschenko comments on "eucal and mathrsfs in jSMath?" (4029)

Anton Geraschenko — Wed, 24 Mar 2010 00:45:53 -0700

@Andrew: I was almost certain you'd say just that :). Like fpqc, I'm also going on the assumption that fonts are only loaded if they are actually used, but I want to test that assumption carefully. I meant to ask you last time this sort of question came up, would it be possible for you to set up a proxy in Norway I could use to browse MO? That way you wouldn't have to keep reminding me about slow connections and far-away places, and I wouldn't have to keep asking you to test if things actually get any slower/faster when I make some change.

Harry Gindi comments on "eucal and mathrsfs in jSMath?" (4024)

Harry Gindi — Tue, 23 Mar 2010 17:50:55 -0700

I'm pretty sure that jsMath does not load the files unless they are used on the page. Anyway, you can just download the proper fonts and not worry about it =p.

Andrew Stacey comments on "eucal and mathrsfs in jSMath?" (4023)

Andrew Stacey — Tue, 23 Mar 2010 17:34:35 -0700

No!!!! Not more stuff to download before I get to see any mathematics!!!

I'm currently on a slow connection in Brazil and it is painful how long it takes an MO page to load.

Harry Gindi comments on "eucal and mathrsfs in jSMath?" (4017)

Harry Gindi — Tue, 23 Mar 2010 03:29:01 -0700

Why don't you just call eusm10 "\eucal" and put a note in the FAQ? I mean, when I think of the font, I think of the package name.

If you'd really prefer it as "\cal", could you set both "\eucal" and "\cal" to eusm10?

Anton Geraschenko comments on "eucal and mathrsfs in jSMath?" (4016)

Anton Geraschenko — Mon, 22 Mar 2010 23:15:10 -0700

jsMath doesn't have a preamble, but there is a way for me to load fonts and define macros. It looks like I'd have to load the rsfs10 font (mathscr) and the eusm10 font (eucal). We don't have those fonts installed right now, and it appears that I don't have write access to latex.mathoverflow.net, so you'll have to wait for Scott to handle it (I think he's out of town for the week). Once we have the fonts installed, I'll define the macros \mathscr and \cal (\mathcal seems to be much more common than \cal for the computer modern calligraphic font, so redefining \cal to be the eucal calligraphic font will be the least jarring).

Harry Gindi comments on "eucal and mathrsfs in jSMath?" (4015)

Harry Gindi — Mon, 22 Mar 2010 18:24:58 -0700

Is there any chance that we could add those two fonts to MO?

It seems like all you would have to do is add the following to the preamble of jsMath:
\usepackage{mathrsfs}
\usepackage{eucal}

This allows us to use any of the three calligraphic fonts by typing in

\CMcal{-} (CMcal - standard)
\mathscr{-} (rsfs)
\mathcal{-} (eucal)

Edit: I looked over the source for eucal.sty again and realized we can just leave eucal as it was (because the .sty defines CMcal).