[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

comments on mathfont-0.5


When I had a look at Matthias Clasens' lastest release of his partial
implementation of a new math font encoding, I noticed the following
problems:

1.  It would be nice if the CHANGES file would use the same date format
used elsehwere in the LaTeX world, i.e. 97/05/14 instead of 97-14-05.
It's just a little confusing that way.

2.  Some of the font tables in mathfont/doc/charts.dvi lie partially
off the page.  The guilty part seems to be the line \textheight=1000pt
at the top of mathfont/tex/charts.tex.  Replacing the 1000pt by a more
reasonable value like 650pt should fix it.

Surprisingly, the comment "%FMi" in the offending line seems to
indicate that the problem came from nfssfont.tex form LaTeX base.  
Is there any good reason for this or simply a LaTeX bug?  (Frank?)

3.  Some of the MF sources in mathfont/ym[a-f] use files with the same
name but different contents.  When all these files are copied to
mathfont/mf and found there first, some of the resulting font tables
will be really messed up.  

One way to resolve this confusion might be to rearrange the whole set 
of MF sources in terms of:

[yma_enc.mf]

  symbol_A:=0;
  symbol_B:=1;
  symbol_C:=2;
  ...

[symbols.mf]

  iff known symbol_A:
  cmchar "Symbol A";
  beginchar(symbol_A,...)
  ...
  enchar;

This way, it would be possible to put all symbols of a certain kind 
in the same source file and have that file included from different
driver files while having only those characters generated that are
specfied in the particular encoding. 

4.  I wonder what happened about my suggestions for rearranging the
MX font table in a more systematic way.  While I understand that 
there is a reason to have all the basic sizes up front, arranging
the extensible sizes in blocks of 8 characters makes the whole 
font table much clearer.   It's just a matter of devising a scheme
and adding \nextslot commands to the .etx files to implement it.
(If I find some time, I might have another try at it myself.)

5.  Concerning the Euler version:  I don't see why you treat Euler
just as another math version.  Shouldn't it actually be another
implementation containing Euler as normal and Euler bold as bold?

6.  Finally, concerning the bold version: While adding this is a step
forward, it's still not enough to satisfy the needs of the physicsts:
We still need a full set of Latin + Greek (i.e. a subset of MC) in
bold sans oblique and/or bold sans upright.  While this shouldn't 
be too difficult for Latin letters, it'll probably need a complete
redesign of lowercase greek to make bold sans greek sufficiently 
distinct form bold italics greek.


So much for now.  I'll have a closer look at the font tables tonight.

Cheers, Ulrik.