On 2002.01.12, at 06:44, Jarkko Hietaniemi wrote:
Well, for that we would need Unicode. How much of it we need?
Is the Unicode::String enough? If you need working Unicode regexes,
then pre-5.8 (yes, 5.8) Perl's are out of luck.
The answer may be that it will be never enough to make everyone
"sufficiently happy" (a ala Camel Book 3rd Ed.). That's one of the
reasons why Encode should just provide API to actual encoders and do
nothing for encoding itself.
It does not mean that Encode module <collection> cannot include
encoders. I just want these features detached from Encode.pm itself.
As for functionality, once again Encode:: and its modules do nothing
more than (en|de)coding and let others like regex engine to do
manipulation. That way you can provide basic encoding capabilities to
old perls while keeping new ones happy.
And there is yet another reason why Encode itself should be made
detachable from perl core. The list of encodings is subject to growing
REGARDLESS of perl core. Chances are that making Encode inseparable
like current implementation makes both perl core and Encode unstable,
like Xwindow and Unix kernel....
Dan the Man with Too Many APIs to learn
