[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Font naming rears its ugly head again
- To: tex-fonts@math.utah.edu
- Subject: Font naming rears its ugly head again
- From: alanje@cogs.susx.ac.uk (Alan Jeffrey)
- Date: Wed, 25 Aug 93 14:39 BST
- Flags: 000000000000
Dear all,
I'm currently working on v1.x of the fontinst package for creating
Virtual Fonts out of AFM or TFM files. A preliminary (very
user-unfriendly!) version was described at the Aston meeting, and is
available from CTAN in fonts/utilities/fontinst.
The fontinst package produces rather different VFs from afm2tfm or
Sebastian's PSNFSS, most noticably in the c&sc fonts, which are
letterspaced.
So, I've been thinking about what I should call these fonts in the
ghastly 8+3 format we all know and... er... love. At the moment, I'm
using Karl's scheme, so the Adobe Times Roman fonts are:
ptmrq Adobe Times Roman u&lc Cork
ptmrcq Adobe Times Roman c&sc Cork
But this isn't totally satisfactory, since it makes dvi files rather
unportable... if you receive a dvi file containing the font ptmrq,
whose ptmrq should you use?
To get round this, what I'd like to do is use a `unique' prefix for
the fonts I generate (say `f1' for `fontinst') and then use my own
scheme for the remaining 6 characters. For example:
f1ptmXXX Adobe Times Roman u&lc Cork, made with fontinst
f1ptmYYY Adobe Times Roman c&sc Cork, made with fontinst
I was planning to use the remaining three characters (after the `f1'
prefix and the 3-letter family name) to encode a 15-bit number, with
something like:
5 bits for weight
1 bit for slanted / upright
1 bit for roman / italic
2 bits for u&lc / c&sc / all-caps / all-small-caps
1 bit for lining / non-lining digits
1 bit for whether it was created with an expert set
4 bits for encoding
Would anyone object to this? I'm sorry to have to break the mold of
Karl's scheme, but I can't see any other way of avoiding clashing font
names...
Alan.