RE: issue-dbooth-9c: Base URI should only be required if needed

Jeremy,

In essence, section 5.1.4 of RFD3986 seems to be saying that it is NOT
an error if a relative URI is used but the XML document has defined no
base URI, because the "context of the application" will then define one
by default.  For GRDDL, it may be more appropriate to consider this an
error condition, but I am willing to accept this as satisfactory
resolution of issue-dbooth-9c (Base URI should only be required if
needed).

However, this does NOT fully address issue-dbooth-9a (GRDDL should be
usable in a messaging pipeline). 

Thanks

David Booth, Ph.D.
HP Software
+1 617 629 8881 office  |  dbooth@hp.com
http://www.hp.com/go/software
 

> -----Original Message-----
> From: Jeremy Carroll [mailto:jjc@hpl.hp.com] 
> Sent: Monday, June 04, 2007 1:27 PM
> To: Booth, David (HP Software - Boston)
> Cc: McBride, Brian
> Subject: Re: issue-dbooth-9c: Base URI should only be 
> required if needed
> 
> 
> PS feel free to respond on-list, particularly if it's of the form, 
> "Jeremy made the following off-list report to me of the discussion at 
> the last telecon, and it adequately addresses my comment" :)
> 
> Jeremy
> 
> 
> Jeremy Carroll wrote:
> > 
> > Hi David
> > 
> > I believe the intent of the WG, at the last telecon, was to 
> > respond to 
> > you on this issue; but I messed up chairing and gave Chime 
> > an action to 
> > respond to you  on issue-dbooth-9a instead, on which we did 
> > not have 
> > adequate consensus, and he correctly bottled out.
> > 
> > I believe the WG consensus at the last telecon was that the 
> > base URI 
> > issue in the case of an XML document for which no obvious URI is 
> > available, is adequately covered by RFC 3986 section 5.1.4:
> > 
> > http://www.apps.ietf.org/rfc/rfc3986.html#sec-5.1.4
> > [[
> >  If none of the conditions described above apply, then the 
> > base URI is 
> > defined by the context of the application. As this definition is 
> > necessarily application-dependent, failing to define a base 
> > URI by using 
> > one of the other methods may result in the same content being 
> > interpreted differently by different types of applications.
> > 
> > A sender of a representation containing relative references is 
> > responsible for ensuring that a base URI for those 
> > references can be 
> > established. Aside from fragment-only references, relative 
> > references 
> > can only be used reliably in situations where the base URI 
> > is well defined.
> > ]]
> > 
> > In the case where there are no relative references, the 
> > above mechanism 
> > uses an application default URI which is then ignored.
> > This is the mechanism that is supported in the Jena I/O API 
> > (for all document formats).
> > 
> > I believe this works adequately in an XML messaging 
> > pipeline, and in the 
> > sorts of applications I believe you are interested in, a 
> > policy of the 
> > sort: "The GRDDL results at each stage should not depend on 
> > relative 
> > references" can be implemented by having an application 
> > default URI, for 
> > example <not-a-known-scheme:///> and then indicating a 
> > policy violation 
> > if any URI in the scheme not-a-known-scheme: appears in any 
> > GRDDL result.
> > 
> > any thoughts?
> > 
> > I guess I would find it helpful to know whether this sort 
> > of response 
> > would adequately address that comment from your point of view, and 
> > whether it also addresses the related 9a issue, that the WG 
> > is still 
> > struggling with. My view is that it does, but then we differ.
> > 
> > Jeremy
> > 
> > 
> > 
> > 
> > 
> 
> -- 
> Hewlett-Packard Limited
> registered Office: Cain Road, Bracknell, Berks RG12 1HN
> Registered No: 690597 England
> 
>

Received on Tuesday, 5 June 2007 20:40:11 UTC