To: mitra@pandora.sf.ca.us
In-Reply-To: mitra@pandora.sf.ca.us's message of Thu, 21 Jul 1994 23:47:59 -0700 <94Jul21.234802pdt.2764@golden.parc.xerox.com>
Subject: Re: Fragment ID;s
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <94Jul21.235707pdt.2760@golden.parc.xerox.com>
Date: Thu, 21 Jul 1994 23:56:57 PDT
I know our messages probably crossed. In fact, in the most recent
draft, I left '#' as 'unsafe' rather than 'reserved'. However, it was
necessary to explain that '#' was used as a separator in a fragment
identifier in World-Wide Web, and that fragments were to be left
inside the optional <> delimiting brackets in free text.
This is treading a fine line of not specifying things that we're not
ready to specify (like 'fragment identifiers') and yet leaving room
for them, by explicitly disallowing URLs with unescaped #.
Many people have pointed out a desire to have some way of designating
fragments of non-HTML documents.
This stuff has been done a zillion different ways in the hypertext
community; I'm certainly uneasy reinventing a design without
re-looking at how previous hypertext systems handled links to images,
sound, etc.