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

Re: revised draft of EuroTeX paper

Two more points; only the first is directly related to the paper (but
it has two parts).


chris

-----------------------------------------------------------------------
The description of .mfd files does not make it clear to which of these
they will set up a mapping:

  the internal LaTeX font selection scheme `name' of the (virtual)
  font;

  the external font file name.

I assume it should be the former, since the .fd files do the rest of
the job.

Also, when referring, as in the .fd paragraph, in the present to the
font selection scheme, please use the term `\LaTeX{} font selection
scheme' (since it is no longer new and there should not be any others
around, it should no longer be called NFSS or NFSS2).  Thanks.


I do not think we should agonise too much about whether or not to
include in a math encoding some particular symbol that is already
available in some reasonably widely available font: eg the Icelandic
letters.

Typically, in my experience, any given document will require at most
one or two such glyphs.  And it is very unlikely that one math
document will use up more than 12 math families.  So a better approach
would be to provide an easy way to set up in a document preamble
math-mode access to a particular slot in a particular encoding and a
font that is set up by the system to be available in LaTeX (possibly
after loading a .fd file).  The user is assumed to know what slot in a
VF is required and either the internal LaTeX specification of the font
or its font-file name.  This is probably not very useful for the lone
user but at a reasonably LaTeX-aware site it is quite realistic for
such information about fonts for use with TeX to be accessible but
that not all are set up for immediate math use in the basic LaTeX
system.


One way to do this is to provide a nice interface to something like:

  \nfss@text {\usefont ... \char ...}

or, eg,

  \mathrel{\nfss@text { ... } }

This has the following features:

--  it does not use up math families;

--  it does not change size unless amsmath is in use (this could be
    fixed by loading code to make \nfss@text do its magic);

--  accent placement will not be so good (only a big problem for slopy
    symbols).


The other way I shall summarise as: internally define a new family (if
necessary) and set up everything needed to access it and the slot
within it.