At 16:52 +0000 2001/02/26, David Carlisle wrote:
>> What characters are included in a set?
>you could look at the tables:-)
Yes I did (apart from the fact that I could find no convenient archive to
pick them down, making the process excruciatingly slow on my computer).
>> If it is a-zA-Z0-9 plus undotted ij,
>> one set has 64 characters, giving room for 1024/64 = 16 sets.
>It varies from set to set. the basic collections are
>letters (a-z A-Z)
>greek (including variant greek forms)
>some of the alphabets don't have greek some don't have grrek or letters.
Looking into the TeX book, there are 40 Greek characters not identical to
Latin. It might be tempting to add a full set of Greek letters, but in
_math_ it seems pointless: letters will mostly appear singly with no other
suitable context information identifying them as Greek. (By contrast, in
Greek text, one will know that they are semantically Greek letters from the
context, and further they may be drawn from special Greek fonts, giving
them a slightly different look from the Latin letters, which may be drawn
from a different Latin font).
If the Greek letters appear in shapes
that gives 160 characters.
This gives at most (1024 - 160)/64, or 13 Latin sets. I think these should be
Sans-serif Bold Italic
with no "Bold Fraktur" and no "Monospace".
-- The monospace is not really a _math_ font, there is no _semantic_
difference in using a monospace over another font, not even when writing
computer language code. So strictly speaking, it is a form of rendering.
And the "Bold Fraktur" seems unnecessary. That is, unless somebody can
demonstrate that it is in actual use.
By contrast, I can think of a (thought) example where Calligraphic and
Script are in use in the same formula: I think the "O" of order O(n) (as in
complexity of algorithms, for example) should be in the RSFS like Script.
But it would be perfectly OK to have Calligraphic letters denoting some
other quantity (say categorical objects, even though some prefer Script for
that too). Well, anyway, one could without too much effort produce sensible
formulas where the two appear side-by-side, indicating different semantic
But perhaps Unicode has already made up its mind, so there is nothing to do