W3C home > Mailing lists > Public > uri@w3.org > December 2002

guidance for choosing a vendor-specific URI scheme name

From: by way of Martin Duerst <mike@skew.org>
Date: Tue, 17 Dec 2002 00:04:48 +0900
Message-Id: <4.2.0.58.J.20021217000442.056f4bb0@localhost>
To: uri@w3.org




Hi,

I'm one of the core developers of 4Suite, a Python-based XML & RDF processing
toolkit and web application development platform. I've been doing a lot of
work recently on bringing 4Suite's URI handling into compliance, and it has
become apparent that we need a custom URL scheme for one part of the software,
so tonight I was reviewing various documents to figure out what's involved in
choosing a name for a new URL scheme for use within 4Suite and in applications
built on it.

The scheme is vendor-specific; if resolution/retrieval is performed, it
implies access via a custom RPC protocol used only in 4Suite. It will
primarily be used only internally, but it will definitely be exposed publicly
and may very well end up in URI references appearing in documents on the web.
I'm not sure if this would qualify as "public" or "private" in RFC 3305
terminology, but I at least want to choose a scheme name that doesn't violate
any standards, whether current or pending.

So far I've looked at

- RFC 2717 and RFC 2718
- RFC 3305 and http://www.w3.org/TR/uri-clarification/
- Internet-Draft draft-king-vnd-urlscheme-02

It looks like a scheme name based on the VND proposal would have been the way
to go, but the Internet-Draft expired on 2001-01-14 and apparently there has
been no RFC formalizing the recommendations it made. I'm wondering how dead to
consider this draft. Is the proposal being taken any further?

Should I choose a "vnd.vendor-ID.foo" scheme name in spite of the lack of
standards? Or can I just go ahead and use shorter name like "ftss"?

Since there's no formally recognized alternative tree (RFC 2717 terminology),
no final RFC governing the VND schemes, and no IANA registration of
Vendor-IDs, there are risks in making the former choice, and the latter choice
is risky just by virtue of its potential collision with something that might
be formally registered someday. However, that doesn't seem to stop anyone
else.

So... what's a would-be URI scheme inventor to do? What's the recommended
procedure in this situation?

Thanks,

Mike

--
   Mike J. Brown   |  http://skew.org/~mike/resume/
   Denver, CO, USA |  http://skew.org/xml/ 
Received on Monday, 16 December 2002 10:26:53 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:31 GMT