On Jan 27, 2010, at 09:44, Glenn Knickerbocker wrote:
> Bob Cronin wrote:
>> suspicions (that things can only get better, unless we run into some legacy
>> application that has built a dependency on the crummy translation done by
>> STANDARD).
>
> That's actually what brought this back to my mind recently: I fixed the
> translation in one FTP application to accommodate the missing "¬" sign
> last year, and recently I got a complaint about having fixed it from
> someone who was using UTF-8 encoding for all data on his Linux machine
> but then using tools that didn't bother to decode it.
>
> That one character is problematic. Since it wasn't the same in all
> extended ASCII code pages (DOS used AA, Unix and Windows used AC) and
> wasn't available on ASCII keyboards, people just got used to thinking of
> "shift-6" on the 3270 and PC keyboards as meaning the same thing and
> expect EBCDIC 5F to translate to ASCII 5E "^" rather than AC "¬". So
> expect complaints either way--but you know that already.
>
What does Rexx accept/support? Programmers who want to write
portable Rexx nowadays use '\' to indicate negation.
'[', ']', '{', and '}' are also problematic. Some say also '|'.
STANDARD is at least better than IGC0010C.
-- gil
