[5055] in www-talk@info.cern.ch
Re: Maintaining HTTP connections...
daemon@ATHENA.MIT.EDU (Bert Bos)
Tue Aug 2 05:33:37 1994
Date: Tue, 2 Aug 1994 11:27:53 +0200
Errors-To: listmaster@www0.cern.ch
Errors-To: listmaster@www0.cern.ch
Reply-To: bert@let.rug.nl
From: Bert Bos <bert@let.rug.nl>
To: Multiple recipients of list <www-talk@www0.cern.ch>
Larry Masinter writes:
|What is (or should be) the relationship between SGML entities and
|URNs? If we had a URN system, would we use URNs for naming the various
|WWW symbols?
I hadn't thought of that, but it's an exciting idea! But I would turn
the second sentence around: many entities are sort of standardized by
the SGML standard, I think they count as URNs.
Some thought needs to go into resolving the words `sort of' above. For
example, the entities are only standard in certain contexts. By mixing
SGML and URL conventions, we could say that an entity reference like
`é' is really a `partial URN', which expands to:
ISO 8879:1986//ENTITIES Added Latin 1//EN#eacute
If we agree on sets of additional entities, we could treat them the
same way, except that the public identifier would be different, e.g.:
-//WWW//ENTITIES WWW standard icons//EN#folder
The question remains: what does the URN reference? The way the
standard entities are used in SGML, they refer only to an abstract
idea, usually a glyph. Depending on the application the abstract glyph
can be turned into dots of ink, sound, a character code in some
charset, etc.
Viewing standard entities as URNs wouldn't change anything in the
browsers, of course. A browser that accidentally tries to do a GET on
such a URN would receive a placeholder, just as the entity definitions
in the SGML standard are just placeholders.
Bert
PS. There would be related URNs that *do* point to real data, such as
the URNs that link to fonts for printing the glyphs.
