- From: Henry Story <henry.story@bblfish.net>
- Date: Fri, 24 Jan 2014 09:11:35 +0100
- To: Alexandre Bertails <bertails@w3.org>
- Cc: Eric Prud'hommeaux <eric@w3.org>, "public-ldp-wg@w3.org Working Group" <public-ldp-wg@w3.org>, Arnaud LeHors <lehors@us.ibm.com>
On 23 Jan 2014, at 23:17, Alexandre Bertails <bertails@w3.org> wrote: > On 01/23/2014 03:58 PM, Eric Prud'hommeaux wrote: >> * Alexandre Bertails <bertails@w3.org> [2014-01-23 15:07-0500] >>> Hi Eric, >>> >>> On 01/22/2014 09:09 PM, Eric Prud'hommeaux wrote: >>>> * Arnaud Le Hors <lehors@us.ibm.com> [2014-01-22 15:36-0800] >>>>> Hi all, >>>>> >>>>> As you all know we've been slipping off track with regard to our schedule >>>>> to deliver a Recommendation by the time our WG expires (1 June 2014). We >>>>> cannot afford any further slippage. See the timeline I laid out: >>>>> http://www.w3.org/2012/ldp/wiki/Meetings:Telecon2014.01.27#2nd_Last_Call_timeline >>>>> >>>>> What this means, practically, is that we no longer have time to >>>>> investigate and debate issues at length. We HAVE to close every remaining >>>>> issue one way or another. If we can't agree on a resolution then we will >>>>> postpone the issue. Yes, that means we may have unresolved issues. Nothing >>>>> is perfect. >>>>> >>>>> The good news is that while we still have several pending issues, I >>>>> believe we can close them all quickly (yes, really). We've got proposals >>>>> for all of them and it's "merely" a matter of getting these accepted. >>>>> >>>>> Of course this is only to get us to LC2 and there is no way to tell what >>>>> will come out of it but that's beyond our control. >>>>> >>>>> In the agenda for next Monday I put forward a set of proposals on how to >>>>> revolve every remaining issue: >>>>> http://www.w3.org/2012/ldp/wiki/Meetings:Telecon2014.01.27 >>>> >>>> Persuant to the agendum for ISSUE-92 - Interaction Model, tests for >>>> creation of sub-container >>>> <https://dvcs.w3.org/hg/ldpwg/raw-file/tip/Test%20Cases/LDP%20Test%20Cases.html#TC-C11> >>>> and archival of container >>>> <https://dvcs.w3.org/hg/ldpwg/raw-file/tip/Test%20Cases/LDP%20Test%20Cases.html#TC-C12> >>>> use these header specifications respectively: >>>> Link: val=<LDP1ContainerInteraction>; rel=profile >>>> Link: val=<LDP1ResourceInteraction>; rel=profile >>>> >>>> Here, <LDP1ResourceInteraction> stands for whatever the WG chooses to >>>> identify the HTTP interactions defined in §5 >>>> <https://dvcs.w3.org/hg/ldpwg/raw-file/tip/ldp.html#h2_ldpr> >>>> >>>> and <LDP1ContainerInteraction> stands for whatever the WG chooses to >>>> identify the HTTP interactions defined in §6 >>>> <https://dvcs.w3.org/hg/ldpwg/raw-file/tip/ldp.html#h2_ldpc> >>>> >>>> These are concrete test cases which I hope will appeal to the folks >>>> waiting in the wings for something implementable. >>> >>> To be clear, the proposal does not change the value for the >>> rel=profile relation, except maybe for the typos in the spec >>> http://www.w3.org/ns/ldp/Container as it's supposed to be >>> http://www.w3.org/ns/ldp#Container :-) Same for ldp:Resource. >>> >>> I know you wanted to avoid the confusion between the class and the >>> object. >> >> make that "interaction model" > > :-) > >> >>> That would actually align with the rationale of changing >>> rel=type to rel=interaction. That being said, I think it's fine to >>> keep ldp:Resource/ldp:Container as values because a class is just a >>> URI as others that can can be used as object. We would just make sure >>> that by dereferencinng them, we found find the information about using >>> them in the context of rel=profile and everything would be just fine. >> >> Defenders of using "ldp:Container" to identify the interaction model >> have stated that doesn't bind us to using only one interaction model >> to manipulate properties of RDF nodes of type ldp:Container. I'm very >> curious how they will support that in the text that comes back when >> you dereference ldp:Container. > > Somewhere in http://www.w3.org/ns/ldp.html, at the fragment-id > #Container, I expect to find something saying that ldp:Container when > used with rel=profile denotes the Container interaction model as > defined in http://www.w3.org/TR/ldp/#the-right-id. You can not do things like that in RDF. You can not have a URI denote one thing if it is related to by one relation, and another thing if it is related to by another relation. > > Somewhere in http://www.w3.org/ns/ldp.{rdf,ttl}, I expect to find a > statement { #Container dcterms:description "@@same thing than in the > HTML@@" }. > > Note: "clients SHOULD NOT indiscriminately access profile URIs." :-) > > Alexandre. > >> I propose something like >> >> "ldp:Container identifies a graph structure with the following schema. >> It does not identify any particular interaction model for manipulating >> nodes with that structure." >> >> >>> Alexandre. >>> >>>> >>>> >>>>> Be prepared to cast a vote on every one of them or accept not to have a >>>>> say. I will not further delay resolution on these. >>>>> >>>>> I know this has been trying for everyone and it's been hard for some to >>>>> stay on top of everything. The thing is I don't think the world would fall >>>>> apart no matter how any of these decisions would go. So, let's just decide >>>>> one way or another and move on. >>>>> >>>>> Thanks. >>>>> -- >>>>> Arnaud Le Hors - Software Standards Architect - IBM Software Group Social Web Architect http://bblfish.net/
Received on Friday, 24 January 2014 08:20:46 UTC