A 05:37 98-04-09 -0700, Keld J|rn Simonsen a écrit :
>> Canadian ordering standard CAN/CSA Z243.4.1 also sorts the whole
>> répertoire. And CAN/CSA Z243.230 defines code-independent LOCALEs (one
>> French, one English) based on this répertoire, both provided with 3
>> charmaps (one for ISO/IEC 8859-1 [one more charmap to be added, the same
>> as 8859-1 plus the extra characters of -15], one for ISO/IEC 6937, and one
>> for the UCS]).
>Oh, I always thought you had 3 repertoiremaps, not 3 charmaps.
>I think the latter makes more sense.
In fact they are 3 répertoiremaps, I did not use the right term as it was
nonly for vulgarization (one needs to consult the standards since there are
many more details -- our full répertoire goes beyond 6937, including a set
of characters defined by Statistics Canada before Latin 1 was made). We do
everything in a code-independent way. The répertoiremaps (a table titled
CHARIDS) define character symbols solely on UCS ids, fitting with
répertoires of many sets of ISO and private codes (Latin 1 répertoire can
be handled in plain ISO/IEC 8859-1, or with IBM 850, or with IBM EBCDIC 037
CECP, or with Unicode). This is one of the 3 répertoires (a 4th to come
with Latin 9, *practically* identical but with extra characters). That is
much more flexible than defining a myriad of charmaps.
Otherwise for Canada, we would need charmaps for all of the above plus IBM
863, plus ISO/IEC 6937, plus NAPLPS, plus *Tetetex*, plus *teletext*, plus
IBM Canadian French EBCDIC (I don't remember the IBM registry numer of this
one), plus... plus... plus... plus subset charmaps... this would be
impractical in Canada, and even more impractical for a standard to refer to
private codes, without denying that in other countries where less diversity
exists, it could make sense to only use actual charmaps based on coding.
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:40 EDT