Alain Fontaine writes...
Hi. I would really like to understand why some people make harsh comments
about the use of ISO646 national variations, while they seem to be perfectly
happy by the (mostly japanese) use of ISO 2022.
In my (probably broken) mind :
- both are somewhat conformant, because only octets in the range 0-127
are used ;
- both are non-conformant, because the real meaning of those octets is not
given by the USASCII table.
Your analysis is correct, and at least some of us are quite intolerant
of both in principle and quite tolerant of both in practice: to
paraphrase a comment made a few days ago, I really don't expect anyone
to speak only English while IETF gets its act together on these issues.
There is one important difference between the two which argues very
slightly for 2022. If one is going to use 2022, the character sets
being shifted to are identified; the only problems are issues of initial
state and "when one goes back" (end of line? end of message?). With
national variations on 646, there is no way to tell which particular
variation is in use. One could, of course, do that with appropriate
header fields, but none of those have been standardized.
With 2022, if a UA "knows" that 2022 might arrive, and is built around
an appropriate profile, the character set is internationally
interchangable. With 646 NLVs (and without a specification of
additional required headers), there is no way to have a UA tell how to
process the characters correctly.
Another possible answer is that you Europeans have been too quiet for
too long. :-)
--john