[lucida] Question about lucida-otf.sty

[Bruno Voisin]

> Le 6 nov. 2017 à 02:11, John H Lienhard <lienhard at mit.edu> a écrit :
> 
> The ongoing development of both unicode-math and lucida-otf.sty leads to some instability, I think. With other unusual features of lucida-otf.sty (math font not scaled by default to match text font; and no initialization of Lucida Grande), I agree that your approach of just making one’s own definitions is more robust until the situation stabilizes.

Two follow-ups:


1) The issue with \mathit etc. not working and defaulting to Latin Modern was recently fixed by Will Robertson (the author of fontspec and unicode-math), following a report I made at

https://github.com/wspr/unicode-math/issues/404

(I had forgotten about that). So it seems, based on your report, that lucida-otf's use of "version=normal" disable sthe fix. Might be worth reporting at <https://github.com/wspr/unicode-math>.

Generally, unicode-math is currently undergoing a lot of changes, and Will's just announced an oncoming release including all existing fixes before big changes later in the year:

https://github.com/wspr/unicode-math/issues/424 


2) By the way, this wasn't the point of your original report if I understood correctly, but just so everything's clear: it's perfectly normal that

\mathrm vs. \symup
\mathit vs. \symit
\mathbf vs. \symbf
\mathsf vs. \symsf
\mathtt vs. \symtt

give different results, these are not the same fonts. The way things are defined by unicode-math:

- Classical math alphabet commands like \mathit switch to the fonts picked by fontspec, such as LucidaBrightOT-Italic.otf.

- The new symbol math alphabet commands like \symit switch to the "Mathematical Alphanumeric Symbols" Unicode block of LucidaBrightMathOT.otf, such as glyphs U+1D44E-U+1D467 for lowercase Roman italic. See pp. 28 ff. of

https://www.tug.org/store/lucida/otmath-blocks.pdf

referenced at the bottom of the TUG Lucida page

https://www.tug.org/store/lucida/opentype.html

(the two links "Lucida OpenType Math by block" and "Lucida OpenType Math-Demi by block").

The glyphs are sometimes different between the two, and the metrical data for them (whichever form this takes in OpenType) are different.

By contrast, since there is no classical \mathbfit math alphabet, all the other old/new-style commands are defined to be equivalent, namely

\mathbfit = \symbfit

and so forth.

The uses of the two styles are different: \mathit etc. for multi-letter Roman symbols, and \symit for single-letter symbols. So in theoretical fluid mechanics, for example, I will use \mathit{Re} for the Reynolds number but \symit{\Gamma} (or the math-style=ISO option) for the Gamma function.

Similarly, to be complete, it may be noted that \boldsymbol from amsmath switches to bold version hence to LucidaBrightMathOT-Demi.otf, while \symbf switches again to blocks U+1D400-U+1D433 (bold Roman), U+1D6A8-U+1D6E1 (bold Greek) and U+1D7EC-U+1D7F5 (bold digits) of LucidaBrightMathOT.otf.

The resulting spacing is subtly different between the two (maybe this is because \boldsymbol introduces some math kern here and there), and the glyphs shapes may be subtly different (but I'm not sure of that).


That's all!

Bruno