Re: Irony heaped on irony

On Fri, 19 May 2000, james anderson wrote:

> > As I said in an
> > earlier email, it's just a name.
> 
> In general, this is correct. A given http-url is not, however, "just a name".

No, but what is at the end of a URL when you access it is undefined - hell
it might not even be static. The namespace spec should not constrain
what's meant to be at the end of a URL, not even for the defined usage of
"identifying a namespace".

> > Bind to your email address if you
> > want. The idea of http URL's was because it provides a simple mechanism
> > for people to get unique names (even someone without their own domain name
> > could use either their personal web space, or geocities or something),
> > rather than using some horrid GUID scheme. The idea had nothing to do with
> > resolving that name.
> 
> If this account is correct, then the notion of using http-url's was
> misconceived. There are plenty of other uri schemas. Many of which
> include domain names. The namespace recommendation should be careful to
> either specify and to use uri's with the appropriate semantic or revise
> the recommended semantic to match the uri which it specifies and uses.

Please go back and read the ancient archives on this. Way back when the
namespace spec came out - which had to come out fast because people needed
the functionality, we started talking about alternative urn schemas. I
don't know what happened to that discussion as I lost track of XML-Dev in
moving jobs.

My point though is just this: Nothing should define what goes at the end
of a URI. If you want to put a schema there, fine. But please don't
anybody start writing parsers that try and see if there's a schema at the
end of the URI - that way lies misery for the whole internet.

-- 
<Matt/>

Fastnet Software Ltd. High Performance Web Specialists
Providing mod_perl, XML, Sybase and Oracle solutions
Email for training and consultancy availability.
http://sergeant.org http://xml.sergeant.org

Received on Sunday, 21 May 2000 09:09:13 UTC