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

Re: More missing glyphs ...

At 21:30 97-04-19, Frank Mittelbach wrote:

>Hans Aberg writes:
> >   The \diamond is listed as a binary operator, whereas the \lozenge is
> > listed as a miscellaneous symbol, so these have different intended use, and
> > the different design seems to reflect that.
>the fact that a glyph has different usage, eg as a binary or as a
>relation or as ...  does not necessarily mean that we have two
>different glyphs or need two different glyphs.
>the fact that glyphs in two different fonts (by different designers)
>look different is also not really a good indication.
>similar examples of the above types are some of the lasy font symbols
>like \rhd (or whatever the are called) which also have counterparts in
>ams fonts and which also do have different status (one being used as
>relation the others as binaries) but still can and should be
>considered one glyph
>it is often a vague boundary and thus cant be fully logically decided
>but the rule of thumb should be that typical look of the glyph and not
>its usage or name should decide

  Just in case somebody wonders, my interpration of Frank Mittlebach's
comments is that he wants to make precise the principles one might use for
classifying those glyphs. (And not as polemics that requires a fight.)

  To me it looked as though the \diamond was designed so it could be used
as a substitute when you run out other similar sized binary relations (like
tho others on page 436 in Knuth's "TeXbook"). It is quite difficult to
invent new such symbols, which are good-looking and also easy to write by
hand, I have noticed, so it is good to have a descent supply of those.

  The \lozenge looks as though it is not as all designed for that useage or
purpose, and I can not immediately recall any particular mathematical
useage. The Merriam-Webster's dictionary states that a lozenge is a
heraldic figure, with equal sides. But the \lozenge appears in the
AMS-fonts package, so those people might know.

  Otherwise, it seems me that there are two different classification
principles in play here: The high-structured, high-level, as in LaTeX,
which deals with intended useage (like writing \tensor for a tensor, or \em
for emphasis), and the low-level, which deals the typical looks (like
writing \otimes for a tensor sign, or \it for italic). Even if you want the
latter to rule the font classification schemes, you have already spent much
effort on classifying math glyphs according to their intended useage (like
having an arrows package, grouping letters together, etc.) Also, often, if
one wants to introduce a new symbol, one often first looks at symbols of
similar useage, and tries to select a similar-looking, but clearly
distinguishable, so the two principles often go hand-in-hand.

> > -- Perhaps one should try to get a compiled list with the symbols intended,
> > or typical, use.
>a very good suggestion, would you think of compiling one?

  I think this could take an effort over a period of time, where different
people would have to report what they know about the life of a particular

  Roger Kehr <kehr@iti.informatik.th-darmstadt.de> mentioned there is a
classification of the international alphabets called UNicode; perhaps the
topic would fit into that scheme.

  Hans Aberg