RE: Use of [ and ] in URIs

I'm concerned about the relationship of XPointer to the
character quoting rules for URIs in general. There might
not be a problem, but I don't see this addressed in particular
in the XPointer document.

# XPointer defines the meaning of the "selector" or
# "fragment identifier" portion of URIs that locate resources 
# of MIME media types "text/xml" and "application/xml".

the URI specification doesn't seem to be listed as a
'related standard'.

What's raised the issue is a proposal (as a result of
work in representing IPv6 addresses in URLs) to move
the characters "[]" into the reserved set. Ron Daniel
noted that XPointer uses within fragments essentially
as reserved characters.

In general, a URL parsing program takes the URL, parses
it using the reserved characters into its component parts,
decodes any URL encoding on the components, and then
interprets the results.

I haven't quite figured out what the operational model
is for XPointer as XPointer strings are moved into
and out of the fragment identifier area.

I think the "Editor's note" 'In other words, the %-escaping
is not really needed when XPointers are used in URIs in
XML documents.' is wrong and should be removed.
If you want to use a syntactic element which allows
characters that are not allowed in URIs, then don't
use a URI. Use something else, e.g., the IURI proposal
in draft-masinter-url-i18n-04.txt. That might change
the content model of XML linking.

Larry
-- 
http://www.parc.xerox.com/masinter

Received on Monday, 4 October 1999 16:03:05 UTC