Re: format nego in HTML/10646?

Martin J Duerst (
Mon, 8 May 95 14:36:10 EDT

>I think this discussion of whether and how to display characters
>from 10646 outside of Latin 1 should be deferred or moved into a
>private forum (perhaps we need a sublist for discussing the I18N
>paper and issues -- I'm sure not everyone wishes to follow this thread).

Probably yes, if we want to get the Internationalization issue
ahead fast. But on the other side, the experience from SGML
experts and others on this list has been a tremendous help.

>What would be useful at this point is to specify in the current 2.0
>RFC the requirements of a UA in rendering character data. I would
>expect these requirements to read something like:

Well, I don't have anything against it, but I get the impression that
what is proposed is more or less what I would expect a reasonable
implementer to do anyway. But maybe I am too familliar with this

I just wanted to mention a point that seems pretty important
to me for actual usability, and that we should solve somehow
if we get into that much detail:

>3. An HTML 2.0 UA which cannot depict a graphic character...
>... should depict such a character as either:
> - a sequence of characters which represent the reference itself
> rather than its replacement text, e.g., ￰

Seen from SGML, the format of numeric character references is
obviously decimal. But unless one is equipped with a calculator
that converts from decimal to hexadecimal (of course assuming
that one has the modern equivalent of an ASCII table in the form
of two Unicode volumes), these references are very tedious to
use in this form. Due to the work involved to decode them,
a "?" character is in my oppinion not worth less.

Please keep this aspect in mind. There are many nice solutions,
and we shouldn't specify behaviour of UA in too much detail
before we know the issues involved.

Regards, Martin.