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

Re: About atomic encoding

Michael Downes wrote:

> > On the other hand, I need various roman letters in upright 
> > type and if they are taken from the operator font rather than 
> > the core font, there are problems with inter-letter spacing.
> > (Just consider an upright `i' followed by `round' letter such
> > as \omega or \sigma, compared to the case when an upright `i'
> > is followed by a 'straight' letter such as \hbar or `k'.)
> Is this really a problem? The general philosophy of TeX fonts with
> regard to math symbols is that adjacent symbols are normally independent
> entities, so they should stand apart a little, rather than being closely
> fitted like the letters in words. In order to better evaluate your
> suggestions, we need to know more about your reasons for rejecting the
> basic approach: making sure that each symbol has enough surrounding
> space built in to keep it from running into an adjacent symbol, and then
> accepting whatever intersymbol spacing results.

The problem I mentioned is based on my observations when using cmr 
(a font with text spacing) for the upright `i' in combination with 
cmmi (a font with math spacing) for the other letters. I guess the 
problem is that the operator font is a text font, meant for multi-
letter expressions like `sin', and it doesn't have math_fitting. 
If the upright `i' would come from a real *math* roman font with 
appropriate math_fitting rather than a text roman font, the problem
probably would go away. This would however require loading another
math symbol font, which would be a waste of families unless that font
can be used for other purposes as well (such as particle symbols).
In any case, the suggestion that physicist could take their upright
letters from the operator font is not a solution to this problem. 

Since the upright `i' and the upright `d' logically are core symbol 
rather than special symbols from a different math alphabet, I'd argue
that it would be preferable to put them in the core font and to allow
implict kerning with other core letters. Besides, kerning between 
`d' and other core letters is specially tested in the \math test 
of Knuth's testfont routine from Appendix H of the METAFONTbook. 
Similar testing would be needed for the upright `d' or `i' as well. 

However, we will have to see how many slots are left after the other
core symbols are allocated. After having looked at Justin's list again, 
I think that after careful evaluation we might eventually get along 
without resorting to multiple core fonts for different disciplines.

Anyway, it's nice to see some discussion about math fonts coming back
again after months of silence...

Ulrik Vieth.