[Python-Dev] Strange error importing a Pickle from 2.7 to 3.2
Stephen J. Turnbull
stephen at xemacs.org
Thu Feb 24 02:14:59 CET 2011
M.-A. Lemburg writes:
> "Latin-1" is short for "Latin Alphabet No. 1" [...].
> I assume that since the HTML standard used the more popular
> name "Latin-1" for its definition of the default character set
> and also made use of the term throughout the spec, it
> became the de-facto standard name for that character set
> at the time.
As usual with de facto standards, it got "embraced and extended".
I've seen people seriously contend that Windows-1252 is an
"implementation" or (conformant) extension of "Latin-1", and that the
EURO SIGN is now a member of "Latin-1". It's just too ambiguous for
my taste; I avoid it in discussions of character sets, preferring to
be thought idiosyncratic and pedantic.
As for the spelling, I think "Latin-1" is slightly more readable than
"Latin1", but the latter is in the same degree more typable.<wink>
> For much the same reasons, "ISO-10646" never really became
> popular, but "Unicode" eventually did.
No, there are much more important reasons why "Unicode" became
popular. IMHO, as an encoding standard ISO-10646 had a slight edge
over Unicode in the early 1990s, before the two were unified as coded
character sets. However, as a text processing system there simply was
no comparison. Unicode provided a large number of standard
facilities, and was clearly set to add to those, that were way outside
of the scope of ISO 10646. Claiming Unicode conformance was a much
bigger deal than ISO 10646 (not to mention having the "advantage" that
you could *optionally* save Intel shorts to disk without swabbing them
first).
More information about the Python-Dev
mailing list