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

[Uwe Lueck]

"Philipp Stephani" <st_philipp at yahoo.de> wrote 19.02.2011 16:07:40:
>Am 19.02.2011 um 13:48 schrieb Heiko Oberdiek:
>
>> plain.tex:
>> | \chardef\@ne=1
>> | \chardef\tw@=2
>> | \chardef\thr@@=3
>> | \chardef\sixt@@n=16
>> 
>> * Why \thr@@ and \sixt@@n have *two* "@"?
>
> Because the @'s replace only one spoken vowel (the long I).
>
>> * \@ne has the "@" at the beginning, not at the end as the others.
>
> Because the "e" is not pronounced as a vowel. The rule seems to be: take one of the syllables, and replace all vowels in it with @'s.
>
>> 
>>> \count@ : count zero
>> 
>> plain.tex:
>> | \countdef\count@=255

same in latex.ltx.

(Also summarizing something from others:) 
* `@' sometimes just is appended or prepended to make the cs private.
  Appending is rather Plain TeX style, prepending LaTeX style, 
  LaTeX just incorporates some Plain TeX without adjusting the style of `@' usage.
* Sometimes it is a compound identifier (e.g. \@let at token)
* It has some resemblance to each of (so sometimes replaces)
  - `a' (ignoring the "monkey tail")
  - `e' (ignoring parts of the "enclosed" `a' above its baseline)
  -`o' and `0' (the digit) (ignoring the inner of the "monkey tail").
\pr at mes: better replace some vowel than ... (even if it doesn't look similar) ...

Cheers :-|

    Uwe