On Jul 21,  8:27, John Cowan wrote:
> RFC 2045 [...] and RFC 1341 [...]:
> # This subset has the important property that it is represented
> # identically in all versions of ISO 646, including US-ASCII, and all
> # characters in the subset are also represented identically in all
> # versions of EBCDIC.
> [...] PEM RFCs (1421 and 1113 respectively)[:]
> # characters which are universally representable at all sites, though
> # not necessarily with the same bit patterns [...]

Only the following characters are equally coded in (virtually) all EBCDIC
variants, and are again equally coded in all ISO 646 variants:
- 52 letters: a b c d e f g h i j k l m n o p q r s t u v w x y z
              A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
- 10 digits:                                  0 1 2 3 4 5 6 7 8 9
- 12 punktuation marks:                 space ' ( ) , - _ / : ; ?
- 7 mathematical & miscllaneous signs:              + < = > % & *
These 81 characters comprise the Syntactic Character Set (CS 640),
in IBM terminology. Generally, these characters can be reliably
transmitted across EBCDIC, and ISO 646 systems, as they will always
be transcoded in the same way. However, space characters tend to be
replaced with tabulator characters, in Unix systems, and trailing
spaces tend to be dropped. Also, some far-east EBCDIC variants have
only 26 (viz. one case of) letters.

> So, I continue to believe that the problems with BITNET are
> produced by the different versions of EBCDIC used by different
> BITNET sites without doing proper transcoding when moving from
> one variant to another.

This was indeed so, until end of 1995, when I ceased to gather
information about EBCDIC. Furthermore, most ASCII-EBCDIC transcoding
tables were based on the same pair of codes, regardless of the
code used locally, so characters not contained in CS 640 were often
distorted on their way to the users' terminals. The EBCDIC side of
this ubiquituos mapping was CP 1047 (or some other number I keep
forgetting) which was not even published before 1991 (I think it was
published only in 1993 ± a year or so). All this was was a complete
mess, often deplored in various IBM user organisations (including my
own presentations).

> For example, when I posted lines containing ">" to the Linguist List
> (a BITNET mailing list), they returned to me containing ")" instead.

This, however, has nothing to do with the ASCII-EBCDIC problem, as
both the Greater-Than Sign and the Right Parenthesis belong to CS 640.

Best wishes,
   Otto Stolz