Lists Home |
Date Index |
On Sat, 2002-02-16 at 11:19, Paul Prescod wrote:
> You're right that it is an uphill battle arguing for fundamental
> architectural change in the most successful information system in
URLs are genuinely successful. The scheme: approach was a good idea.
However, I don't think any strong case can be made that URIs are genuine
contributors to the success of that information system, except to the
extent that they overlap with URLs - and, in many ways, damage the
usefulness of URLs.
> URIs are primarily a conversational black hole. In terms of actual
> practice they work quite well. I have had few practical problems with
In actual practice I find URIs work quite well because treating them as
URLs usually works.
In the cases where URIs look like they should exhibit the typical
properties of URLs but don't (this week's case was SAX2 features and
properties in an automated conversion of the O'Reilly SAX2 book), people
get very confused, and quickly.
In cases where URIs look nothing like URLs, there are fewer cases of
such confusion but also a lot less immediately practical that can be
done with them.
There is _some_ substance to the Emperor's clothes, but I suspect that
the merely "conversational" black hole you see suggests that there is in
fact a real black hole underneath. The proponents of URIs have mostly
left readers dazed rather than enlightened, and I don't see
enlightenment spreading. Claiming the success of URLs as the success of
URIs does little to help the situation.
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!