Re: Mozilla 1.9 MathML status and notes - non-Unicode fonts

On Sat, 02 Feb 2008 15:52:18 -0500, William F. Hammond wrote:

> 1. Clarification: Among all the font collections previously recommended
>    at projects/mathml/fonts you are recommending uninstalling only those
>    in the "deceitful font" category, as described in Bugzilla/400938.
>    That is, the other (categories 2 and 3) non-unicode fonts will be OK.
>    Is this correct?

I was recommending uninstalling fonts in the "Deceitful Fonts"
category.  They are probably only going to cause problems if a
document explicitly requests these fonts, so it may not be
_necessary_ to uninstall them.

Category 2 "PUA-only Unicode fonts" won't cause a problem.

Category 3 "Fonts with a Microsoft Symbol charmap" may cause
similar problems.  Different platforms treat these fonts
differently, and there is disagreement about what Unicode points
the glyphs in these characters should be mapped to.

https://bugzilla.mozilla.org/show_bug.cgi?id=399636

These fonts have a Macintosh Roman cmap as well as the "Unicode,
non-standard" cmap, and Linux (and maybe Mac) use this cmap.
The problem is the glyphs are not representative of the code
points in the Roman cmap.

Even on MS Windows which uses the "Unicode, non-standard" cmap,
many applications (or maybe its the OS libraries) map the PUA
Unicode points to the Latin Unicode block (which results in a
mapping slightly different to that obtained by mapping the
Macintosh Roman charset to Unicode).

http://www.microsoft.com/typography/otspec/recom.htm#sym

At least with these Microsoft Symbol fonts, we could detect them,
and choose a behavior.  I have my thoughts on what the appropriate
Unicode mappings should be, but the problem is that there is more
than one opinion (or expectation) on this.

I don't expect or suggest that MS Windows should uninstall their
SymbolMT font, so my best advice is the web authors should not
request the "Symbol" font family.