[texhax] "@" : vowel or glottal stop ? (was : Some puzzling TeX)

[Philipp Stephani]

Am 24.02.2011 14:23, schrieb Paul Isambert:
> Le 24/02/2011 00:09, Martin Schröder a écrit :
>> 2011/2/23 Uwe Lueck<uwe.lueck at web.de>:
>>> broadcast by River Valley. The script appears in TUGboat
>> Shouldn't that be a TUGbo@ ? :-)
>>
>> I'm wondering: Has anybody ever asked DEK about the choice of @ in one
>> of his Q&A's?
> 
> As you might have noticed, I'm not DEK. However, here are some
> reasonable hypotheses about the choice of @:
> 
> Given ASCII, alternatives are:
> - $ and _, but they're already in use for other purposes; same for #, &...;

This didn't stop the L3 team from using _ and : internally, or the
ConTeXt team from using ! and ?. As long as you make sure that internal
macros never bleed into areas they don't belong (e.g. by having only
protected macros), there is no problem with turning any (for exceptions
see below) character into a letter.

> - The hyphen - is out of the question otherwise dimension assignements
> would be impossible;

Right, same with , . and digits.

> - ? and ! are already used in ConTeXt, and probably DEK wanted to make a
> difference;

A difference to what? TeX was obviously invented after ConTeXt.

> - The vertical bar | would have been nice, but then its role in
> Indo-European graphemics is far from clear.
> - Hence @. QED.

What about / ( ) [ ] < > * ~ ` : ?