Suggestion: Expand Stylistic Set 2

[grivasgervilla]

Hi, first of all, I want to congratulate you on this fantastic font which I will use in my PhD thesis. I think that it is one of the best options in the LaTeX environment, and I hope that this project continues to evolve.

I would like to make a suggestion for the next version of STIX2Math font. I think that could be useful to expand the Stylistic Set 2, so a different (lowercase and uppercase) Latin alphabet is used in math mode, instead of the STIX2-Italic font. This could make the inline math in a text more clear and non-content-dependent (like can occur when use mathpazo with eulervm math font which has a different Latin alphabet).

Many thanks for this project.

[tiroj]

There seems to be a lot of support for this idea, but I am not sure what such a variant alphabet should look like.

[grivasgervilla]

Hi, I just want to share an example of the combination of eulervm font with mathpazo in LaTeX:

I do not know whether it could help you to take some ideas about the new alphabet design. The idea is that the math font can be distinguished from the non-math font italics or not, with a different alphabet consistent with the rest of the STIX2Text font.

Many thanks for your time.

[davidmjones]

I could bring this up with STI Pub, but personally I find it hard to be enthusiastic about it. Focusing solely on the pragmatic issues, what is the scope of the required changes? The Euler fonts don't just replace math italic a-z and A-Z; they also include harmonized digits, punctuation, and greek, script, and fraktur alphabets. For a Unicode font, that would imply duplicating most of the Mathematical Alphanumeric Symbols, plus the Letterlike Symbols and Number Forms.

If you want your text and math fonts to be more distinct, I think it would be simpler to choose a different text font. In my opinion, expecting a single font to support two very different designs is asking a lot.

[grivasgervilla]

@davidmjones, you could be right, I am not a font designer. I was thinking of some new alphabet on the same line stylistic set 2 changes the symbols for g, u, v, w, and z.

[davidmjones]

@davidmjones, you could be right, I am not a font designer.

Just to be clear, neither am I. I trust that @tiroj will correct me if I say anything too absurd.

I was thinking of some new alphabet on the same line stylistic set 2 changes the symbols for g, u, v, w, and z.

My understanding of the current ss02 set is that it represents common off-the-shelf types of customizations of individual letter shapes -- single-storey vs. double-storey g, for example -- that co-exist within the overall design of the alphabet. (Or maybe that exist separately from the overall design? As I said, I'm not a font designer :)). I know there is also a single- and double-storey a, but I don't know how far this generailizes to other letters.

On the one hand, I'd be open to including any other common typographical variants of this sort (such as the double-storey a), but on the other hand, I don't know if that really accomplishes the stated goal of making math and text more clearly distinguishable. But then again, I'm not even sure making them more distinct is a good idea, so it's hard for me to judge.

[tiroj]

Stylistic Set 2 in the Math font does a couple of different things:

a) replaces the italic default double-storey g in the STIX Two design (derived from Times New Roman) with the hooked alternate form (also applies to math sans style);

b) replaces the special rounded forms of u, v, and w that are the default math italic style with the angular forms that are the default in the STIX Two Text Italic (as noted previously, this set does not extend to the bold italic or ssty and ssty2 variants, which is something we might revise in 2.20 if we have time).

So with regard to the u, v and w, that feature is actually restoring the STIX Two Text italic forms in the Math font, whereas with regard to g is it providing a variant form.

I am not entirely happy with shoehorning these two different functions into a single stylistic set feature, but we had to be economical because there are only twenty such features to work with, and a lot of variant sets in STIX Two.

---

This issue seems to be suggesting something more radical, especially if the eulervm example is taken as a model. That package seems to marry the upright forms of the original Euler math fonts with a version of Palatino (also by Hermann Zapf, but not part of the original Euler project). Obviously this does produce a strong distinction between text and equation because they’re different typefaces.

[mdeff]

For what it's worth (I'm not a designer either), I am of the opinion that math and text should blend seamlessly. As suggested by @davidmjones and @tiroj, if one wants them distinguished, one can change font.

Without changing font, it's impossible for the unstyled (i.e., not italic, bold, sans, script, fraktur, etc.) math alphabet to be different from the text alphabet, because they are the same code points.