Dear all,
I really did not want to open a can of worms. I just wanted
some guide lines on how we might name our namespaces and also if we
chose http:// based ones that we might have a home to which the URL
pointed.
The message I am hearing is that it is a can of worms and that
we are probably best to resove it at TC level. As to the home fo rthe
schema files etc I think I will take it up with Scott and see if there
can be a space outside KAVI for some of this info or a URL rediction
tool that can work for http://oasis-open.org/tc/tcame/....
As to the use of versioning. Well I really did not expect the
debate I expected a dictated solution from the Oasis board as I feal
this kind of thing is part of the brand as in the W3C, but hey I am sure
people will now say that even they are inconsistent:)
Martin Roberts
xml designer,
BT Exact
e-mail:
tel: +44(0) 1473 609785 clickdial
fax: +44(0) 1473 609834
Intranet Site :http://twiki.btlabs.bt.co.uk/twiki
-----Original Message-----
From: Eve L. Maler [mailto:]
Sent: 24 February 2004 15:50
To: Norman Walsh
Cc: Eduardo Gutentag; Roberts,MME,Martin,XSG3 R;
; ;
Subject: Re: OASIS Naming Rules
(Norm and I have had this discussion before...)
Norman Walsh wrote:
> From a software point of view, changing:
>
> urn:oasis:names:tc:SAML:2.0:assertion
>
> to
>
> urn:oasis:names:tc:SAML:2.1:assertion
>
> has exactly the same semantics as changing it to
>
> urn:blorty:warble:brrr:its:cold:outside
>
> There's no way to preserve existing semantics if you do that. Many
> languages (HTML, DocBook, XSLT, ...) expect V2.1 of the language to be
> "mostly the same as" V2.0 so they use the same namespace name and
> transmit version information in some other way. (Or not at all.)
Yes, but... Many languages declare their intentions ahead of time for
what constitutes a "namespace version upgrade", which gives you the
ability to predict tighter semantics for an n.m to n.m+1 change or an
n.m to an n+1.0 change. SAML is one; it will stick with n.0 until a
major backwards-compatible revision is done, and anything in that
namespace that *hasn't* been changed will have the same semantics as it
used to.
And if you register a URN namespace foo, you have the right to
officially declare that urn:foo:...:... has subfields that mean
particular things, and thus you can make such intentions formal.
Eve
--
Eve Maler +1 781 442 3190
Sun Microsystems cell +1 781 354 9441
Web Products, Technologies, and Standards eve.maler @ sun.com
