OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Namespaces. Re: [xml-dev] RDDL (was RE: [xml-dev] Negotiate Out TheNoise

[ Lists Home | Date Index | Thread Index ]

----- Original Message ----- 
From: "Bullard, Claude L (Len)" <clbullar@ingr.com>

This letter brings us one year back.

> Doesn't this essentially say:
> 1.  If you use http:// it is a URL. Period.  In effect, 
> there is no real difference between a URI and a URL 
> that uses http.  The expectations of the web gospel 
> are different in the new testament.

If you use http:// it is a URL. It looks like URL - 
it should be URL!  

So in fact, from the beginning,  namespaces should 
not be using  http:// prefix, because by design 
namespaces were not URLs but just a hidden 
unique properties, attached to each and 
every tag. Writing it in form of http:// was a 
mistake of W3C and RDDL speculates on that 
mistake, playing it to the bone.

Abusing Namespace URI ( hidden, 
semantic-neutral 'just be unique' property) 
with any linking functionality is plain stupid and starts 
influencing almost every existing XML tool, as RDDL shows.
( 'added support for RDDL to XSV')  for the purposes
that are *unclear* to those who are 'maintaning' 
the RDDL 'technology'.

RDDL is a plague. They're playing with the virus, 
not understanding what they really do. The virus 
should be placed under control of some solid 
organization, like W3C, because it influences 
each and every XML document and each and 
every XML tool  in the world. Yes, that's it. 

W3C should step in and either crush or adopt RDDL
until it is too late.

> 2.  The namespace specification should be rewritten 
> to include the possibility of resolution of namespace 
> identifiers and state that when http is there, expect 
> it to be resolved, so require a document and suggest 
> some examples of what might be expected, plus a default 
> if none (eeeeee... better than 404?) so then RDDL might 
> become a W3C spec and the default.

God forbid. Namespaces specification should 
explicitely say that URLs should never be resolved 
by any software in a sense other than being unique 
hidden properties of every tag,  until W3C specifies the 
way those URIs / URLs  should be resolved and 
any software that makes any assumption that there 
is some particular resource at the end of namespace 
URI is a violation of XML specification.

To avoid confusion, W3C could change it's 
Namespaceing naming policy and start writing 
Namespaces URIs without the http:// prefix, 
so that people will not try using these URIs.

However,  because of legacy problems, 
better solution could be that the namespaces URLs 
would start pointing to some human-readable 
documentation, related to the namespace, so that 
people would see that when some namespace is 
written as an URL it is reasonable to expect some
documentation at the end of that URL and *that 
is the only purpose of that URL*.

No processing software should use Namespaces
URLs for fetching *any* data. Wanna fetch something 
from the web? Use other technologies, don't abuse the 
namespaces URI.

> I can't tell if this opens up a can of worms or 
> if the fish aren't going bite.

The can stays opened for more than one year. 
RDDL just made it worse.



News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS