FW: OASIS Naming Rules

[<martin.me.roberts@bt.com>]

From: Eduardo Gutentag [mailto:Eduardo.Gutentag@Sun.COM] 
Sent: 24 February 2004 15:05
To: Eve L. Maler
Cc: Roberts,MME,Martin,XSG3 R; Norman.Walsh@Sun.COM;
scott.mcgrath@oasis-open.org; ebxml-bp@lists.oasis-open.org;
cam@lists.oasis-open.org
Subject: Re: OASIS Naming Rules


It would probably be a bad idea if the TAB Quality document on naming
conventions ever tried to impose normatively whether namespaces should
or should not carry versioning information. Personally (and I believe
I'm not alone) I believe that kind of decision rests squarely with the
TC. In particular, it appears that some governmental agencies insist on
the use of the version attribute, rather than reliance on namespaces, so
the normative imposition of a namespancing usage one way or another
would be totally out of line. And who invented the use of namespaces for
this purpose anyway? ;)



Eve L. Maler wrote:
> Hello again Martin,
> 
> I'm not the final word on this; the OASIS TAB's quality subgroup will
> be.  But in advance of hearing from them through Eduardo (note that
it's 
> a holiday in the U.S. today), I'll give you my opinions for what
they're 
> worth.  I think the SAML TC has done a pretty good job of getting 
> through this thicket, and it seems to reflect some growing consensus 
> about the "right thing to do," so I'll use it as an example.
> 
> URNs have become favored by quite a few people and organizations for
> their namespace URIs.  The SAML TC uses them for its two (so far) 
> namespaces, and it does include version information:
> 
> urn:oasis:names:tc:SAML:2.0:assertion
> urn:oasis:names:tc:SAML:2.0:protocol
> 
> If you look at the draft SAML V2.0 spec (Section 4, SAML Versioning),
> you can see our explanation of how we handle versioning in general:
> 
> http://www.oasis-open.org/committees/download.php/5232/sstc-saml-core-
> 2.0-draft-04-diff.pdf
> 
> 
> Because we relegate (just about) all our backwards-incompatible 
> changes
> to major versions, our XML namespaces don't change for minor versions,

> though we do set the "version" attribute on <xs:schema> to reflect the

> specification suite version in which the schema document was
published.
> 
> Regarding resolution of namespace URIs, we in the SAML TC haven't set 
> up
> any catalog-like way of resolving URNs.  We assume that schemaLocation

> hints will be needed.  I'm noticing that in SAML V1.1, we filled in
the 
> schemaLocations with URLs, but we've gone back and forth on whether to

> put a relative filename there.  If you have experiences to share, I'd 
> love to hear them.
> 
> Note that SAML also uses URI-based identifiers extensively for other
> purposes (see Section 7 in the same document).  Again, we put versions

> on these that match the original specification suite version in which 
> the URI was first published.  If the semantics of any one identifier 
> were to change significantly, I'm sure we would update the version in 
> the URI.
> 
> I hope these thoughts help,
> 
>     Eve
> 
> martin.me.roberts@bt.com wrote:
> 
>> Eve,
>>     I am working on two TC schemas where the naming of the schemas 
>> namespaces and locations of schemas.  We could do with this being 
>> resolve quickly.
>>
>>     The two issues we need resolving, is what should an OASIS 
>> namespace look like - for example should be be versioned, should it 
>> be a URN -urn:org.oasis-open:tcname:schemaname:version or 
>> http://oasis-open.org/schemas/tcname/schemaname/version.xsd.
>>     In both of the above, but especially in the second it would be 
>> good to know if the url could be resolved to a particular location.
>>
>>     Thanks.
>>
>> Martin Roberts xml designer, BT Exact
>> e-mail: martin.me.roberts@bt.com tel: +44(0) 1473 609785  clickdial
>> fax: +44(0) 1473 609834
>> Intranet Site :http://twiki.btlabs.bt.co.uk/twiki