[XeTeX] Fractions using fontspec

Will Robertson wspr81 at gmail.com
Thu Mar 12 07:23:38 CET 2009


On 2009-03-12 16:00:28 +1030, "R (Chandra) Chandrasekhar" 
<chandra at ee.uwa.edu.au> said:

> I am using TeXLive 2008 from the DVD and kept updated regularly. I am
> trying to get fractions using the [Fractions=On] feature in fontspec.

I think the fractions feature has been labelled "troublesome" and it's 
recommended not to use it. I'll have to address this in the fontspec 
documentation at some stage. Oh wait; I might be getting mixed up with 
the "Ordinal" feature. Erm, Jonathan or Adam?

I don't have these two fonts so I can't investigate in detail, but I 
might have some answers below.

> 1. Why is the digit 1 in Lining style or uppercase in each of the
> plainly written fractions?

Not sure...it might be a conflict between the font and XeTeX's handling 
of the fraction feature; like I said above, best to avoid it.

> 2. Why does the digit 4 in the \vfrac{} version look different from the
> digit in the plainly written fraction, but only in the roman family?

I suspect that \vfrac is actually doing the right thing here and 
selecting the correct glyph; because it's been drawn for the smaller 
size it looks different.

Try this to see the difference:

    \textsubscript{1234}\par\textsubscript*{1234}

(and/or with \textsuperscript as well.)

In the first case, you should get the "proper" subscript numerals; in 
the second case, simply a rescaling down of the originals. (Although 
the situation might be more complicated if there are optical sizes as 
well.)

> 3. Is the a quirk of the version of the Adobe Minion Pro and Myriad Pro
> I am using?

I think everything looks like it should, give or take.

> 4. What is the preferred method of getting fractions using solidus
> notation in XeLaTeX for OTF fonts?

Do you mean by using a solidus in the source? I'm afraid (as far as I 
know) that there is no robust solution across all fonts, since it 
depends on how the font is constructed. Your best choice is to use 
\vfrac.

Cheers,
Will




More information about the XeTeX mailing list