From: Doug Ewell (firstname.lastname@example.org)
Date: Thu Feb 05 2004 - 03:16:34 EST
Dean Snyder <dean dot snyder at jhu dot edu> wrote:
>>> I hope Apple re-thinks this, because it makes PUA useless in plain
>> That's because it is. Without further specification, the PUA is
>> completely ambigious.
> So you're saying PUA is only useful for rich or marked-up text?
There has to be an understood convention for interpreting the code
points. This could be indicated by some form of markup within the text,
or it could be an unwritten assumption. Of course, this is really no
different from the out-of-band information needed to interpret any
character data -- you have to have *some* basis for assuming that text
is Unicode instead of EUC-JP, or ASCII instead of EBCDIC.
I'll bet some people will be uncomfortable with the notion that a
convention for interpreting PUA code points can be an unwritten
assumption. There is no standard mechanism for indicating how PUA code
points are to be interpreted, so I might have to *assume* that any PUA
text that comes from Dean Snyder's Web site uses XYZ Cuneiform
convention, any that comes from the Klingon Language Institute's site
(or mine) uses ConScript assignments for Klingon (or Ewellic), and any
that comes from William Overington's site uses the interpretations that
that gentleman sees fit.
My pages that include PUA code points contain a short, human-readable
HTML comment explaining that such a convention exists and giving a URL
where more information can be obtained, but who knows if that is the
right way to do it? There is no standard.
Dean may not be interchanging his PUA Cuneiform text with anyone else,
but there needs to be a convention between him and his operating system.
In contexts where he can choose his font freely, he has the right to
pick the font that will display his PUA code points the way he wants
them. In others, he's out of luck. On Windows, I can't even rely on
being able to display real Unicode characters for Vietnamese in places
like the Start menu or the title bar of the browser, because they're not
in the one and only font used for each of those places.
This archive was generated by hypermail 2.1.5 : Thu Feb 05 2004 - 06:19:54 EST