# No Subject

• To: math-font-discuss@cogs.susx.ac.uk
• From: Michael Downes <MJD@MATH.AMS.ORG>
• Date: 05 Aug 1993 10:38:01 -0400 (EDT)
• Cc: MJD@MATH.AMS.ORG

My 2 cents worth.

> 2 - I believe that NFSS is a major improvement over the
> default scheme used in LaTeX 2.09. However, it leaves a lot
> to be desired for. Namely:
>
>       NFSS assigns *all* the math fonts at each *size* command.
...
>       The solution to these problems is to load math fonts *truely* on
>       demand, that is *within* each math formula.
...
>       be replaced by new commands such that, for example, \Gamma
>       could be defined as \MathCharDef\Gamma7\rm00 .
>       This means that the \rm command would load the default
>       font for \Gamma and then typeset the 0'th character in the font.

In order for this argument to be convincing you must explain how you
handle loading of the proper math font for \Gamma if it occurs in
\scriptstyle or \scriptscriptstyle. Also the loss of speed that you
mentioned later is a compensating factor that weakens your statement
that NFSS leaves a lot to be desired'.

>       One could also modify \MathCharDef so that these symbols can be used
>       both in math and in text modes freely.

The primary reason why \mathchardef'd characters produce an error
message outside of math is to help the user catch mistakes that
involve forgotten begin-math or end-math commands. If you use
a different \MathCharDef that gives no error message the user loses
this aid.

> 3 - I have often found myself wondering whether the
> "design size"  approach has lived out its usefulness.

I support Martin's response, that specifically designed smaller size
fonts for subscripts provide better readability, regardless of the
resolution. AMS editorial staff and mathematician authors who have no
knowledge of the technical details underlying math font usage have
noticed and complained about unsatisfactory appearance of scaled-down
fonts used in subscripts. Indeed it is fairly obvious to the eye, I
would say, if you compare.

> Besides, I should also have mentioned that AMS now uses Times-Roman as
> its main font] in its journals. If they see it acceptable to scale
> fonts, who am I to argue?

The use of scaled Times Roman math roman and math italic fonts at the
AMS is not an intentional abandonment of the design size approach but
an expedient because it was desired to use Times Roman but the
phototypesetter manufacturer (Autologic) did not provide specifically
designed subscript sizes of the font. Given an arbitrary 10-point
scalable font in PostScript or other format, I would say that the
ideal way to tailor it for math usage would be to add specifically
designed smaller sizes, at least, let's say, for the 7 and 5-point
sizes as Blue Sky and Y&Y have done in their PostScript versions of
the CM fonts. But that is an expensive and time-consuming task. Given
a choice between devoting some hundreds or thousands of staff-hours to
font development or accepting lower quality, it's obvious which way
most publishers and other users will choose; but that doesn't mean
the users won't recognize and admit the loss of quality.

> There is one more major problem though. And that is the huge size
> of the PostScript files generated when one uses different
> design sizes, which in many cases yields a VM error on most modest
> printers.

Sorry, this is irrelevant. If the current technology for some
PostScript printers has difficulty supporting the design size
approach, that does not change the fact that the design size approach
provides higher quality typesetting. Your original proposal was that
the design size approach should be abandoned because you believed it
didn't give any significant quality advantage.

[PS. Justin, I think the design-size discussion is not completely
irrelevant to the math font encoding discussion, since practical
limitations on the number of different fonts that can be loaded in TeX
or printed may suggest that the math font encoding should be organized
along the lines of grouping together the symbols most commonly used,
rather than along the lines of common design features---in order to
allow most users to get by with (let's say) 6 or 7 mathgroups instead
of 11 or 12.]

> I hope this rather long piece of "complaints" will be viewed
> as a constructive, albeit possibly naive, attempt to stimulate
> discussion.

I would say it's always valuable to challenge the assumptions, to make
sure that they are reliable and weed out ones that are not.

Michael Downes                              mjd@math.ams.org (Internet)

`