RE: [xml-dev] Namespace URI question

From
Jeff Lowery <>
To
"'J.Pietschmann'" <>, Jeff Lowery <>
Date
2002-12-12T20:59:00Z
ID
<>
Thread
RE: [xml-dev] Namespace URI question
> From: J.Pietschmann [mailto:]
> > xmlns:foo="file:///foodom.xsd"    

> Not quite correct.
> The relevant RFC is
>   http://www.rfc-editor.org/rfc/rfc2396.txt
> (unless it has been obsoleted by a newer one recently.
> Your URL is an absolute URL. 

Okay. So the above is an absolute file URL to nowhere. 

>>  Meaning its an absolute path,
>> relative to the root-of-the-moment.
> I have no idea what this should mean.

It means I misunderstood the meaning of the URI.

> In any case, it is unusual and perhaps unwise to use a file URL
> to indentify a name space. While the spec is fuzzy enough to allow
> basically arbitrary Unicode strings, the rationale for choosing
> URIs is to uniquely identify the name space, i.e. to provide a
> string unique in the whole universe. Using HTTP URLs comes
> reasonably close to this intention.

Yep, I agree. The example I chose did not originate from me, and my
arguments for not doing this were the same as yours. But I was a bit
confused on the interpretation of file:///file.ext URI format. I see it now
as a rootless file URL, and therefore meaningless. I doesn't appear to be
officially deprecated for namespace usage, although it perhaps should be. I
can certainly see a lazy programmer using this for a short identifier, but
unwise URIs similar to this may be too to numerous to enumerate for
deprecation.