To: uri@bunyip.com
Subject: are #anchor-id's allowed in Gopher or FTP URLs?
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <94Jul1.111234pdt.2760@golden.parc.xerox.com>
Date: Fri, 1 Jul 1994 11:12:33 PDT
This is actually current practice, isn't it? You can say
ftp://host/dir1/dir2/name.html#FOOT1
and get to the FOOT1 component of name.html, can't you?
Also with Gopher URLs that point to .html documents?
The current specification doesn't allow #fragments within gopher and
FTP urls, and that seems to violate both what's widely implemented and
what is ultimately useful.
While it's fine to insist that each scheme allow its own definition of
which of =@;/#?: are reserved in each scheme and which must be URL
encoded, I'm less sure that we should break existing practice in the
pursuit of that.
Can HTML documents occur in news articles? Should news: URLs also
allow # anchor identifiers?
Honestly, I'm not trying to introduce last minute changes, I'm just
trying to figure out how to write the document in such a way that it
is clear what characters have to be encoded and which ones do not.
The March 1993 draft clearly allows # anchor-IDs in any document.
I actually don't remember any discussion where it was decided to
disallow them in gopher and FTP urls.
Maybe the way to resolve this is to make "#" unsafe, take anchor IDs
out of this document, and relegate anchors to the same place that
partial URLs went: to the WWW document for which this should be a
subset.