[texhax] Math operators

[Michael Barr]

Although I don't know anything definitive (if there is anything 
definitive, which I doubt) the following rule appears to be in effect--and 
in any case, it is the one I use.  Multi-character identifiers are set 
upright and single character identifiers are in mit.  This is based more 
on practice (where the differential operator d and the base of the natural 
logarithm e are generally seen in mit).  It makes sense since that is how 
we distinguish s*i*n from sin (there is also a thin space after sin).

The old rule seems to have been that constants are upright and variables 
in mit, but where does that leave the differential operator d?  In any 
case the distinction is not always clearcut.  For example, the sentence, 
"\pi can be as small as 2 on a sphere" makes perfect sense.

It seems to have been computer scientists who started using (at least on a 
large scale) multi-letter identifiers for variables.  Now at least some 
mathematicians are doing likewise.  Putting them in mit would only sow 
confusion.  I guess you could use \textit, but the distinction between 
that and mit is mostly in the spacing and pretty subtle.

The distinction above is the one I enforce (more or less) as TeX editor of 
an online journal (Theory and Applications of Categories).  It is purely 
pragmatic, not based on any principle.

Michael Barr