[texhax] help with identifying some macros

[Barbara Beeton]

On Tue, 20 Oct 2009, Victor Ivrii wrote:

    On Tue, Oct 20, 2009 at 7:15 AM, Dan Hatton <vi5u0-texhax at yahoo.co.uk> wrote:
    > On Tue, 20 Oct 2009, Lars Madsen wrote:
    >
    >> I would rename them into something like \union and \intersection, X \cap Y
    >> does not make any sense.

    Well then one will need to rename all their derivatives (and
    pseudoderivatives) as well including \Cup and  \sqdoublecap or maybe
    even \Coffeecup and \Capricorn (should we?)
    Sure if we had a time machine we could give better names from the very
    beginning but now we can only create a royal mess and pain

the particular names in question were
assigned by don knuth.  this is only a
hypothesis, but i believe that he was
following a long tradition of proofreaders
to use the names \cup and \cap.  if this
is so, then i seriously doubt that any
change would be be likely, particularly
in light of the alternate meanings that
have been pointed out.

it might be noted that in fact there
*are* some alternate names for some
symbols, either because they are in
different classes:
  \triangle, \bigtriangleup
  \backslash, \setminus
or for some other reason:
  \ne, \neq
  \ni, \owns
  \wedge, \land
  \vee, \lor
  \neg, \lnot

this list obviously favors knuth's own
needs, but it shows that there is no
built-in prejudice against multiple
names.  my guess is that \ne and \ni
are so similar that the ease of finding
typos in a file where both are used
would be more important than ease of
typing, and for the logical operators,
the clarity of meaning would be the
deciding factor for which was used.
special-purpose packages have renamed
quite a few symbols for the sake of
clarified meaning (i'm thinking here
of the zed package).  the important
thing is to avoid re-use of names
that have already been assigned.

but authors seem to prefer ease of
typing to clarity of content.  this
wreaks havoc in the ams production
system, where fully linked, on-line
posting of journal articles requires
use of hyperref -- which redefines a
very large number of the "short names".
so these all have to be changed by hand,
leading to the possibility of introduced
errors.  no amount of warnings against
not redefining one- or two-letter names
seems to have any effect.  (this has
gotten off the original topic, for
which i apologize, but if anyone has
a compelling argument that will convince
authors that "clarity is better than
brevity", i'll be delighted to learn it.
i also apologize to authors with limited
manual capability, but there are other
ways to get around that, like editors
with abbreviation-expansion features.)
						-- bb