Re: Best Practices for Establishing Namespace Name

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

C. M. Sperberg-McQueen writes:

> The record would not be complete without someone observing
> that it also has the disadvantage that if the domain name
> registration for [your committee] ever lapses, there is no
> guarantee that the new owner will refrain from assigning
> a new meaning to http://[your committee]/namespaces/xxx

For sure.

But be careful using this as an argument in favour of using URNs or
some other URI scheme -- If you don't care about dereferencing, it
doesn't matter that http://[your committee]/namespaces/xxx doesn't
dereference today, and might dereference misleadingly in some
relatively unlikely future.  If you _do_ care, then the mechanisms put
in place to support dereferencing of e.g. URNs will almost certainly
be vulnerable to the same human-originating failure modes.

> The TAG is presumably aware of this problem (having had it
> pointed out to them more than once), and has presumably
> decided that it does not matter enough to make a difference.
> But as long as domain names are allowed to be used more
> than once by different owners, no system of unique
> identifiers built on the domain name system actually
> guarantees uniqueness of identifier.

Correct.  I am unaware of _any_ examples of published
standards-related URIs transitioning from useful to misleading (as
opposed to useless).  It's not impossible for this to happen, but
pretty unlikely.

I have a long-standing interest in coming up with some kind of
domain-name insurance for web/internet standards . . .

> In the short run, that doesn't matter.  Whether it matters
> in the long run depends on how long you want things to
> work, and how dependent you want to be on netscape.com,
> for example, continuing to denote what it once denoted.

Jonathan Rees has argued, convincingly for me, that
e.g. http://www.w3.org/1999/xhtml will always _denote_ the XHTML
namespace, by use and precedent, _regardless_ of whether or not the
w3.org website persists, or indeed even if w3.org gets reassigned to
someone who starts serving pornographicly embellished caricatures of
the XHTML spec. editors from that URI. . .

I don't expect everyone to be convinced on _that_ point, but what _is_
true I think is that "thinks [will] work" wrt e.g. that namespace URI
regardless of who owns the domain or what they put at it, because
software, unlike human beings, does _not_ depend on successful
retrieval from namespace URIs for proper functioning.

Which brings us to Eliot's point: indeed the reason for the bit about
dereferencability he quoted is that you _can't_ depend on retrieval,
because namespace owners are _not_ required to support it.

ht 
- -- 
       Henry S. Thompson, School of Informatics, University of Edinburgh
                         Half-time member of W3C Team
      10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
                Fax: (44) 131 651-1426, e-mail: ht@inf.ed.ac.uk
                       URL: http://www.ltg.ed.ac.uk/~ht/
[mail really from me _always_ has this .sig -- mail without it is forged spam]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)

iD8DBQFKnpHzkjnJixAXWBoRApUkAJ4vG8p93NLkk1SUQulx09y2QQFh3gCfSfuo
2qr8GNAr5ony8g47zs0CyhU=
=vaad
-----END PGP SIGNATURE-----

Received on Wednesday, 2 September 2009 15:41:28 UTC