Lists Home |
Date Index |
>Unicode has arrived to kill off all of the short sighted legacy
>character encodings and while unicode has a *lot* of problems for asian
>languages (Han unification was *NOT* a good idea), it remains
>infinitely better than the tower of Babel we had before.
I think you definition of the word 'encoding' is a little too broad. This
is not about encodings. It's about giving users the freedom to choose which
parts of Unicode they want. How to specify the method to convert character
numbers to binary form is another problem, and I'm happy with the solution
This way, those who need to use characters in the intervals forbidden in
XML 1.0 would have the freedom to use them, while the rest of us are left
If I'd decide, there would be no change in XML. But if a new version is
unavoidable and I need to pick one, I'd rather go for a more flexible
solution, because I fear that 1.1 won't be the last version of its kind.
I admit there may be better ways than using a PI. Maybe the information
about legal characters is better to specify at the schema level rather than
for each document, but these things are just details. The main point is
that specifying which characters are legal in the specification itself may
be too limiting. Perhaps the only real requirement should be that the
character numbers have the meaning specified in Unicode?