Re: Excess URI schemes considered harmful

At 05:41 PM 9/25/01 -0500, Dan Connolly wrote:
[...]
>I suggest that putting the year-of-issue in a URI
>(e.g. http://www.w3.org/1999/xhtml) carries
>this signal more effectively than putting urn: at the beginning.
>
>Hmm... the idea of having names issued by consensus is
>an interesting one. Surely we could do that inside http
>space, without deploying any new URI schemes, no?

Well, that's a reasonable viewpoint.  I'd point out that urn: isn't a new 
scheme, and already has a defined consensus process for allocation.  I tend 
to think that effective use of urns might, in the longer term, help to 
limit the profusion of URI schemes.

[...]
>Hmm... I wonder how these marketing people
>would feel if the IT part of the company
>changed all their phone numbers after they put them at the
>bottom of a big press release.
>
>404s hurt in tel: space just as much, if not more, than
>404s in http: space.

True.  But even a mature labelling scheme like telephone numbers, where the 
costs are widely-understood, gets changed from time-to-time (area code 
changes, companies moving premises, etc.).

[...]
>This "apparently arbitrary" rule is not arbitrary. That
>it is a fundamental element of web architecture is a reflection
>of the fact that it matters in real life to real people
>who spend real money.

Again, true.  But I think that users of the technology who don't need to 
understand the architecture will relate the cost issues to visible causes 
(like the cost of reprinting brochures).  The practice of periodically 
reorganizing company web sites is with us, and I think it isn't about to go 
away.

To me, it seems easier to let HTTP URLs be HTTP URLs.  Organizations like 
W3C choose provide additional stability for their URLs, but requiring all 
organizations will adopt such a view seems to me like an uphill struggle.

#g


------------
Graham Klyne
GK@NineByNine.org

Received on Wednesday, 26 September 2001 06:55:30 UTC