[that is, all the characters in the ISO 8859-1 coded character set]
> and the
>decision has already been made to provide canonical entity names.
[I take it you mean the entity sets in the appendixes of the SGML
spec]
> It
>doesn't make sense to do both of those things and then fail to provide
>entity names for an arbitrary subset of Latin-1.
..
>This is not an I18N issue. We are not talking about expanding
>Latin-1; we are talking about supporting it consistently.
Right. My mistake. This is an issue of internal consistency in
the DTD.
> there is no logically consistent position on this but to
>standardize entity names for all of the 191 graphic characters in the
>Latin-1 spec. Either that or leave them out.
I agree that the HTML 2.0 DTD is inconistent in this respect. It's
also inconsistent in that there's a URN attribute on A and not on IMG,
and probably in several other ways.
But the HTML 2.0 standardization effort is not about fixing HTML, so
much as describing it as is. Technical changes -- especially those
that affect conformance of widely deployed user agents -- at this
point are MOST unwelcome, at least to me.
I could be convinced to make this change, but it would take something
like all the HTML user agent vendors promising to support these entity
names in a release within the next 6 weeks -- that or a good whack on
the head from the chair of the working group ;-) I'm pretty firm on
this. Good thing I'm not the chair, huh?
If you want to propose a change to HTML, I suggest you write a new
internet draft.
Dan