[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [xri] Nested XRDS documents in Ref processing (was RE: [xri] Special XRI TC Telecon 1PM PT Thursday 2007-10-01)
Steve, Ahhh, my fault, I didn't understand your original message. I thought you were saying that the CID for *foo had to be a child of @ootao*steve. I just talked with John, and since you and Les both feel strongly that the context of the XRDs representing the rest of the authority resolution chain should be preserved, I believe the rest of the TC will support that. It also makes the ref= attribute on the nested XRDS simpler as it will be an exact match of the value of the Ref attribute that produced the nested XRDS. John also suggested what I think is the best solution to tracking a Redirect in the XRDS document, which is to nest it as an XRDS document just like with a Ref, except use a redirect= attribute on the XRDS container. This is one step more complex that what Markus would prefer like (in his email this morning, he'd prefer a Redirect to have direct substitution semantics, so you don't see the original XRDS at all). However it would provide the "audit trail" that Wil and Les felt strongly about, so I expect we can reach consensus on that. It also harmonizes how both Redirects and Refs are represented inside an XRDS document. I will proceed immediately to write all this up in revision #4 of the proposal page at http://wiki.oasis-open.org/xri/XriCd02/RefProcessing. I'll send email as soon as it's updated. =Drummond > -----Original Message----- > From: Steven Churchill [mailto:steven.churchill@xdi.org] > Sent: Tuesday, October 02, 2007 3:57 AM > To: xri@lists.oasis-open.org > Cc: 'John Bradley'; 'Kermit Snelson'; andy.dale@ootao.com > Subject: RE: [xri] Nested XRDS documents in Ref processing (was RE: [xri] > Special XRI TC Telecon 1PM PT Thursday 2007-10-01) > > > There was never any question that the final XRD in the nested XRDS is the > parent of *foo. It has been this way for at least 18 months. > > > XRDS: { > > XRD: *ootao > > XRDS: { > > XRD: *steve > > XRD: *steve's Ref for $res*auth > > } > > XRD: *foo > > } > > By moving the XRD for *foo into the nested XRDS you've done nothing more > that erase contextual information (ie. the ref traversal structure) from > the > audit trail. > > ~ Steve > > -----Original Message----- > From: Drummond Reed [mailto:drummond.reed@cordance.net] > Sent: Tuesday, October 02, 2007 12:25 AM > To: 'Steven Churchill'; xri@lists.oasis-open.org > Cc: 'John Bradley'; 'Kermit Snelson'; andy.dale@ootao.com > Subject: [xri] Nested XRDS documents in Ref processing (was RE: [xri] > Special XRI TC Telecon 1PM PT Thursday 2007-10-01) > > Steve, we discussed the question you asked again on the special telecon > today. Our conclusion is that the example on > http://wiki.oasis-open.org/xri/XriCd02/RefProcessing is correct, even > though > as you point out it is a change from what had been in earlier drafts. > (Note > also that two errors in the example were corrected during today's call -- > the Ref element was in the wrong place and the value of the ref= attribute > was incorrect.) > > The reason for enforcing that all XRDs are in the nested XRDS document is > to > preserve the proper chain of authority for CanonicalID verification. To > use > the example at > http://wiki.oasis-open.org/xri/XriCd02/RefProcessing#head- > 9d7bccc6f42688cc94 > 6ac8412539d751872e8d50 for reference (since it has the full detail), it > shows that when resolving @a*b*c, when @a*b has a Ref to @x*y, from that > point on, the CanonicalID verification chain is rooted in @x. Only @x can > assert a CanonicalID for @x*y, and only @x*y can assert a CanonicalID for > @x*y*c. > > So in essence, when a Ref appears during the authority resolution phase, > it > means that a resolver MUST substitute a new XRI for the old XRI being > resolved, and thus that the only verifiable CanonicalID that can be > returned > is for the new XRI. (This has been written up in more detailed in the > revisions to the proposal at > http://wiki.oasis-open.org/xri/XriCd02/RefProcessing that I made tonight). > In the example above, that means @x*y*c is substituted for @a*b*c, and > @x*y*c is the CanonicalID that the resolver can verify because it > traverses > the entire chain with a Ref, and thereby can confirm that each authority > is > authoritative for its child. > > Note that @x*y*c could still assert @a*b*c as its CanonicalEquivID, and > the > resolver could verify this in its final step by looking for an EquivID > element with the value @a*b*c. > > Hope this helps, > > =Drummond > > > -----Original Message----- > > From: Steven Churchill [mailto:steven.churchill@xdi.org] > > Sent: Monday, October 01, 2007 12:59 PM > > To: xri@lists.oasis-open.org > > Cc: 'John Bradley'; 'Kermit Snelson'; andy.dale@ootao.com > > Subject: RE: [xri] Special XRI TC Telecon 1PM PT Thursday 2007-10-01 > > > > > > I'm not clear about the "Nested XRDS Documents" wording at the > > RefProcessing > > page on the wiki. > > > > For @ootao*steve*foo, say *steve's $res*auth service has a Ref. > > > > XRDS: { > > XRD: *ootao > > XRDS: { > > XRD: *steve > > XRD: *steve's Ref for $res*auth > > } > > XRD: *foo > > } > > > > In this case then another XRD would follow the nested XRDS, no? > > > > ~ Steve > > > > > > > > > > -----Original Message----- > > From: Drummond Reed [mailto:drummond.reed@cordance.net] > > Sent: Monday, October 01, 2007 12:06 AM > > To: xri@lists.oasis-open.org > > Cc: 'John Bradley'; 'Kermit Snelson'; andy.dale@ootao.com > > Subject: [xri] Special XRI TC Telecon 1PM PT Thursday 2007-10-01 > > > > Following is the agenda for a SPECIAL unofficial telecon of the XRI TC > at: > > > > Date: Monday, 01 October 2007 USA > > Time: 1:00PM - 2:30PM Pacific Time > > > > Event Description: > > Special telecon to discuss XRI resolution reference processing as > > described > > below. > > > > TO ACCESS THE AUDIO CONFERENCE: > > Dial In Number: 571-434-5750 > > Conference ID: 5474 > > > > > > AGENDA > > > > 1) REVISED REF PROCESSING PROPOSAL > > At last Thursday's XRI/XDI TC telecon we had extensive discussion of Ref > > processing. We were not able to close on the telecon, however subsequent > > discussion has produced an updated proposal posted tonight (Sunday) at: > > > > http://wiki.oasis-open.org/xri/XriCd02/RefProcessing > > > > The three key changes in this new proposal (vs. the one discussed on the > > TC > > call last Thursday) are: > > > > a) Addition of a Redirect element to take over the function of what was > > previously a "service ref" (which Steve points out operated differently > > than > > an "authority ref"). > > > > b) Keeping the refs= parameter as a boolean (as it was in ED05). > > > > c) Elimination of the proposed "find all" option -- too complex to be > > introduced at this stage (and can be performed by a separate utility if > > needed). > > > > PLEASE REVIEW THIS UPDATED PROPOSAL BEFORE THE CALL IF YOU HAVE TIME -- > it > > is of paramount importance that we close on ref processing so we can > > complete ED06 and proceed to a Committee Draft vote.
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]