At 16:47 -0400 2008-07-24, Jimmy Wales wrote:
Michael, can you share with us what is wrong with
parabaik, myanmar3, padauk?
They may all be compliant with Unicode 5.1. They probably are. They
may or may not have support for minority-langauge characters or for
some special Sanskrit-language shaping behaviour (important for
Buddhist terminology.) I have proposed that we initiate a test suite
to determine what the capabilities of each font are.
I want to make sure I understand something....
If we set the encoding on the site to utf-8, then anyone can use any
myanmar Unicode 5.1 - compliant font to look at it.
No. That's not enough, because a user could look at a page with a
Unicode 4.1 font -- and get an unreadable result. Or a user could
look at a pseudo-Unicode font like Zawgyi (and there are others)
This is in the situation in English, right? Wikipedia
sends me
bytes in utf-8, and then my browser can render it with any normal
font... ariel, times roman, comic, whatever I please.
That works because all the fonts have the same encoding.
So, what difference is there in this case? If there
are Unicode 5.1
fonts, perhaps incomplete, perhaps complete, then fine. People can
use those. What would the problem be with this? Would it render
the site unusable for people who are using only Zawgyi?
We don't want people creating text using Unicode 4.1 fonts or Zawgyi
or other non-conformant fonts. We want text to be conformant,
interchangeable, cut-and-pasteable, and so on.
Zawgyi was a solution, but its users must migrate. There are costs to
that, but we can't support Zawgyi. (We could convert its glyphs and
encoding to 5.1 conformance for people who like it, in principle.)
--
Michael Everson *
http://www.evertype.com