6

I've started exploring XeTeX and the unicode-math package in order to use unicode in my input.

Of the six math fonts described in unimath-symbols.pdf, Latin Modern Math seems to be closest to what pdflatex produces. However, I've already noticed a number of differences I don't like, such as \varnothing (which is the same as \emptyset now), \complement and the \mathbb family.

I know about the range option of the \setmathfont command. Right now I use:

\setmathfont{latinmodern-math.otf}
\setmathfont[range={"2100-"214F,"2201,"2205,"1D7D8-"1D7E1,"1D538-"1D56B}]{xits-math.otf}

But I'd rather use a single command, option or package that takes all symbols as close as possible to the `pdflatex versions'. I can then explore the different fonts at my leisure, with the certainty that there are no big surprises in my existing documents.

Is there a way to do this?

mhelvens
  • 6,126
  • 1
    The appearance of symbols is decided by the font designer. – egreg Jul 23 '13 at 17:35
  • For most symbols, yes. But perhaps you misunderstand the question. --- Which font does pdflatex use for each symbol? And is there a simple way to get XeTeX to use the same fonts in the same situations? – mhelvens Jul 23 '13 at 17:40
  • 2
    With the traditional setup, \varnothing and \complement are taken from the AMS symbol font; the designers of Latin Modern Math had different ideas about those symbols. – egreg Jul 23 '13 at 17:47
  • Well, that accounts for two differences. :-) You wouldn't happen to know about a comprehensive list of such differences? – mhelvens Jul 23 '13 at 18:03
  • I can offer some code for displaying all symbols in a math font, but this would be an answer to a different question. – egreg Jul 23 '13 at 19:00
  • No, that wouldn't really help. But thanks. --- For now, I'll just take the differences one at a time as I encounter them. In fact, after fixing those few I mentioned in my question, there don't seem to be many left. --- I'll keep the question open, though. If someone gives an answer, it might prove useful to others. – mhelvens Jul 23 '13 at 19:12
  • 4
    This question appears to be off-topic because no reasonable answer can be provided. – yo' Sep 06 '14 at 21:32
  • 2
    @tohecz: I'm sorry, but is there any way in which this is not silly? [1] The question is obviously about LaTeX and follows the rules, so: on topic. [2] A reasonable answer to "Is there a way to do this?" might be "No.". [3] How does the inability to answer a question make that question off-topic anyway? --- I can understand the impulse to close what appears to be a dead-end question, but please find a better justification. – mhelvens Sep 07 '14 at 21:12
  • @mhelvens Please complain to StackExchange powers. I didn't come with the word "off-topic" here. Off-topic is a bad notion that is supposed to cover the cases when the question somehow doesn't "fit in". The question really appears to be impossible to answer, which makes it unsuitable for SE, or "off-topic", as the powers above call it. – yo' Sep 07 '14 at 21:18
  • 3
    @tohecz: Why is "No." not a legitimate answer here? Anyway, I asked the question to solve a real problem I faced at the time, and it literally follows all the rules listed in the 'Asking' section of the help center (I checked). It is not subjective. Moreover, it addresses an issue that people may yet face in the future. If you are hung up on the word "answerable" in the 'what to avoid asking' section, I find that a bit weak. The most you could say is that this question does not have an answer yet. – mhelvens Sep 07 '14 at 21:31
  • Well, I think this is a one for meta. I mean, if you open a discussion over this in meta, if you open it on meta, I'll support discussing it. I'm not 100% convinced that it should be closed and that the close reason is reasonable. – yo' Sep 07 '14 at 21:34
  • @tohecz: Truth be told, I don't personally need an answer to this one any longer. But I do find the decision to close/put-on-hold an illogical one. Plus, the question may yet be useful to others. So for the sake of tex.se (and to oppose illogic) I will open that discussion on meta. Though forgive me if I let it wait until tomorrow (yawn). --- All the best, – mhelvens Sep 07 '14 at 21:44
  • @mhelvens No worries. I hope you're not angry at me for casting the first vote and that you don't take it personally since it's not like that! G'night. – yo' Sep 07 '14 at 22:07
  • @tohecz: Oh, please. Don't worry about that. You're just trying to clean the place up. :-) – mhelvens Sep 08 '14 at 10:40
  • 2
    @egreg -- if the question hasn't been asked yet whether there is code to display all symbols in a math font, i'm willing to ask it. the news that latin modern has elided \emptyset and \varnothing is very upsetting to me -- i had a long argument with the unicode folks (which i lost) about whether these were different, or at least different enough to be recognized as legitimate variants. so i need to do some more research to find out whether such "merges" would affect ams publications. – barbara beeton Sep 13 '14 at 20:12
  • @barbarabeeton In my opinion, \emptyset and \varnothing should be a stylistic variant of each other: I can't think to a document using both with different meaning. – egreg Sep 13 '14 at 21:03
  • 1
    @egreg -- i agree with that opinion, but was unable to persuade the unicode folks that the "variation selector 1" (or "VS1" = FE00) should be assigned for that purpose. (the symbols to which VS1 has been assigned are described in unicode technical report 25, "unicode support for mathematics", in table 2.9, p.28.) i guess i'll try again. – barbara beeton Sep 13 '14 at 21:19
  • Looks like that this question is a combination of calligraphic + blackboard bold + empty set symbol. I can't find one for complement – user202729 Sep 11 '21 at 15:46
  • For collection, there can be other differences as well regarding delimiter size https://tex.stackexchange.com/questions/647491/unicode-math-delimiters – user202729 Jun 12 '22 at 09:50
  • and https://tex.stackexchange.com/q/120065/250119 for mathcal/mathscr – user202729 Jan 21 '23 at 11:55

0 Answers0