Re: issue-base-param

Chimezie Ogbuji wrote:

> It is worth noting that RFC 2396 gives us guidance on the 30X scenario and in particular supports (for instance)
>  Jeremey's interpretation of relative URI references in http://purl.org/NET/erdf/ as requiring a resolution against
>  http://research.talis.com/2005/erdf/
> 
> 5.1.3. Base URI from the Retrieval URI:
> [[[
> Note that if the retrieval was the
> result of a redirected request, the last URI used (i.e., that which
> resulted in the actual retrieval of the document) is the base URI.
> ]]] 
> 

Thank you, that's helpful.
The quoted text is also in RFC 3986, also section 5.1.3


Looking through the rest of your message, I think we are in agreement 
except about:

a) HTML base vs XML Base when they conflict

b) Whether we are reopening the issue-base-param

on (b) since, I don't particularly disagree with the resolution, I do 
not require a reopening, but we do seem to agree that we need more text 
to clarify which base param is used in which context; and we need text 
to clarify our relationship, if any, with xml:base. De facto, that seems 
to me like reopening the issue.

On (a) I will try and build a test case before todays meeting, that I 
think is a slam dunk (is that the phrase? I hope it isn't rude!)

More on (b): an issue is that this is the second issue during the last 
call, where we have found all our implementations to be in error.
Particularly since Chime has found the text that demonstrates the error, 
this means that formally:
+ changing the spec to highlight the need to respect redirects, and
+ adding a test to our test suite that has a redirect and the correct 
answer,
is not a design change, but an editorial correction, given the 
misreadings of the earlier document.
However, it gets increasingly implausible to skip CR.
It feels to me that we should be pointing out to the HTML WGs and other 
SW WGs some of the clarifications that we have been making. It could be 
argued that the appropriate process would be a second last call; or we 
could have a CR.


>> Do either library functions:
>>    embeddedRDF
>> and
>>    glean-profile
>> respect xml:base and/or html base?
> 
> As XSLT 1.0 transforms they are not *required* to respect xml:base, but
> they do have to respect an application-provided notion of the "base URI
> of the document entity" - which our resolution provides.  They *can*
> choose to explicitly match xml:base (or HTML <base>) attributes (as the
> RDFa transform does).  

As part of our deliverables they should be consistent with the rest of 
our spec. Thus, if we decide to depend on xml:base, I would be inclined 
to add xml:base support to embeddedRDF, ditto with glean-profile - the 
implementation of rel="profileTransformation" should mimic the behaviour 
in the spec of rel="transformation".

Jeremy

-- 
Hewlett-Packard Limited
registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Wednesday, 18 April 2007 09:11:02 UTC