>> I have also done
>> some work on designing a global character repository, and a global
>> glyph repository architecture to get around the limitations inherent
>> in even a very large finite coded character set.
>That's interesting. Can you say more about what kind of data the
>character repository would contain?
I have been thinking of something conceptually like like the TEI
WSD's. This is basically a description of the abstract character. I
imagine being able to use that in conjunction with the global glyph
server to provide the capability to (in effect), build coded character
sets and fonts on the fly (kind of like the Adobe CID idea).
The global glyph server would have a number of different
naming/indexing schemes, such as code position of a coded character
set, FPI, character name, etc.
>> I do not feel comfortable prescribing much in the way of behavioural
>> characteristics in what is essentially a data fromat description. As
>> such, if one does want to prescribe behaviour in cases where fonts are
>> not available, it should be done through a seperate document.
>Would you like to remove section 2.3 completely?
Not entirely. I think what we have is sufficient.
>Of course the HTML/i18n can be made a "bare bones" specification
>and the explanations and implementation advice be put into
>another internet draft.
Yes, this is the idea, thoguh I wonder if a draft is applicable given
W3C's claim to control over the HTML specification.