Re: HTML 2.0 LAST CALL: Hyperlinking, Forms, Elements

Fri, 2 Jun 95 02:19:57 EDT


| (1) at least one of the proposed names, ¢ isn't supported by
| Netscape 1.0N (just now tested it). I doubt it's supported by _any_
| browser in existence, perhaps excepting the SGML browsers like
| Panorama. Putting this list in the HTML 2.0 spec would require
| changes (albeit small ones) in existing user agents.
| (2) all the entity names in the current html 2.0 DTD are supported by
| all the widely deployed browsers. (I used to test this with netscape
| and NCSA Mosaic and a few others on a regular basis. I believe other
| folks have tested the old ISOLatin1.html against lots of other
| browsers. And the lack of any complaints on this matter from the
| 10,000 folks that used the HTML validation service has got to count
| for something.)
| (3) the current list of entity names has not changed since it was
| first published as an internet draft in June of 1993.
| (4) this list was supported in mid-1994 by NCSA Mosaic 2.4, upon which
| almost all subsequent browser development has rested (at least from
| the point of view of interoperability).

Hmm. OK, I see where you're coming from. But I still say that it's
staggeringly inconsistent to specify entities for most of the
supported character set and leave arbitrary holes here and there in
the sequence.

You seemed to be suggesting earlier that you were thinking of filling
in the gaps in 2.1. But there really isn't going to be any clean,
modular way to do that. There is absolutely nothing controversial
about the list of entity names (except whether we should be using them
in the first place, which is a dead issue). Given that we already
know where we're going, I would rather get it over with than implement
it in lumpy stages that are going to leave different parts of the
sequence lying around in different specifications under different
public identifiers.