Re: Standardization of TeX names for Adobe PostScript fonts.

[Don Hosek <DHOSEK@HMCVAX.CLAREMONT.EDU>]

-re: Proposal for standardization of TeX names for Adobe PostScript fonts.

-Given below are the six character abbreviated file names used by Adobe for
-their outline fonts. It seems reasonable to simply adopt these short names
-when refering to Adobe PostScript fonts in TeX (which happens to be limited
-to six character font names).

-Need for standardization of names for Adobe fonts used in TeX seems to be
-the most urgent, since these are the fonts most frequently used (in fact, I
-have yet to see an author call for a font other than CM, LaTeX, AMS or Adobe).

Oh, there are some of us using PS fonts from other sources
(Cassady & Greene) or non-PS fonts (Bitstream) in daily work. The
reason you don't see any calls for help in my case, at least is
that I have things well under control. 

-Using the vendor supplied abbreviation has clear advantages:

Yes, but the short term view will _always_ lead to big problems.
There already lists a canonical list of TeX abbreviations for I
believe nearly all of the Adobe library. Addressing your specific
points:

-(*) There is no need for a committee to dream up abbreviations for new fonts. 

I dream up the abbreviations I need as I go along. For example,
last week I determined that fosr would refer to the Fluent laser
fonts [Cassay&Greene] Odessa Script font.

-(*) There is no need for a clearing house to approve proposed abbreviations.

Until the inevitable time that somebody discovers that they have
a conflict between the metrics for Times on their HP LaserJet III
and the Times on the typesetter that their final output is going
to. Or that the new dingbats font they bought wants to use the
same file name as their Garamond font. Don't think short term.
Besides, approving proposed abbreviations is simply a matter of
checking to see if their already exists an abbreviation for the
font or if the proposed abbreviation is already in use. The
latter could be done by a computer program, the former could as
well, I suppose, but since there is a need to be wary of
misspellings, abbreviations, etc. this could be difficult.

-(*) There is no delay between publication of an outline font and  availability
-    of a standard abbreviated name for it.

Until the times mentioned above.

-(*) The probability of confusion is reduced when only one short name needs
-    to be remembered for a font (instead of the one used by the vendor AND
-    one approved for use in TeX).

How much word processing/DTP/digital typography software do you
use? I have seen very little software that uses the abbreviated
names in the end-user interface. Even in the case of TeX, the
abbreviated names should only come into play when designing a
style option for using the font under lfonts.new. Also, people
tend to stick to a single application for dealing with printing
of this sort anyway so even if the short font names were used at
any time other than installation, chances are they'd only come in
contact with one version anyway.

-There are two possible problems with my proposal:

-(.) Other vendors may use a particular abbreviation for different font.  

Pretty big

-(.) The present scheme allows for only (26 + 10) * (26 * 10) = 1296 font
-    families.  Adobe already has about 770 / 4 = 193 font families. So
-    in the distant future they are going to run out of two letter combinations.

Let's see, Adobe has been around for around 10 years. We'll give
them an average of 20 families per year. They have 1100 font
families left, so we're fine until roughly 2046. I think we might
be rid of the eight character restriction by then. I certainly
*hope* MS-DOS will have died out by 2046.

-(.) In TeX there is sometimes a need to remap the encoding of the font.  TFM
-    files for the remapped versions must be distinguishable from the `raw'
-    versions. 

Karl Berry's scheme provides for this.

-The first problem can be fixed by prefixing these names with a code
-for the vendor - as suggested by Karl Berry - perhaps `p' for Adobe.

-The solution of the second problem is to simply adopt whatever scheme the
-vendor comes up with when that happens (and this won't happen for quite a
-few years anyway...).

-The third problem is less of an issue now that TeX can handle character sets
-with 256 characters.  But in any case, a suffix can be added to the name 
-(perhaps `x') to indicate remapping (although this does not tell one HOW
-the font has been remapped).  This, along with the vendor prefix, brings the
-maximum length of a name to 8 characters, which almost all file systems now
-are able to deal with.

At this point your proposal is giving names that are logistically
very much like Karl Berry's names. So why not just use Berry's
scheme? It was not generated in a vacuum. There were quite a few
of us who made suggestions on how to map names; the scheme has
already been adapted by one dvi-to-ps system (Rokicki's dvips)
and is likely to be adapted by others. Your scheme doesn't
approach what to do about vendors who (a) don't supply short
names (they exist; there are fonts sold only to the Macintosh
community which can, nevertheless, be converted to a more generic
PS format) or (b) have short names that are already 8 characters
long (Cassady & Greene). Not to mention that Bitstream's short
names are catalog numbers! Shall I continue to call Bitstream
Dutch Roman "11"? I hope not.

In short, I don't think that Vendor-supplied names are adequate
in the least (and I say this from experience with quite a few
font vendors). Adobe is *not* representative of the rest of the
typeface world, so basing assumptions on what can or should be
done on what Adobe does is a mistake.

-dh