- From: Nandana Mihindukulasooriya <nmihindu@fi.upm.es>
- Date: Fri, 10 Oct 2014 11:30:59 +0200
- To: public-ldp-comments@w3.org, Henry Story <henry.story@bblfish.net>
- Message-ID: <CAAOEr1m_xYPj1jYVOfRTCofqO=FWQOvFx3qqQkS9xvyjPXOjjA@mail.gmail.com>
Yes. I think it is important to prevent any accidental overwriting of data. This is addressed by the spec if the server requires conditional requests for existing resources as suggested by 4.2.4.5 [1], In that case, these accidencial overwrite is prevented because any new resource representation without an etag will result in a 428 (Precondition Required). However, if the server doesn't require precondition, I guess the only way for the client to know this would be to do a HEAD in advance but it is not safe because it could always happen that some other client creates the resource at a point between the HEAD and the PUT. Another point related to using PUT for creation. Based on the resolution of ISSUE-91 [2], the interaction is driven by the rel='type' Link header. IMHO, it implies that when creating resources using PUT clients should do the same by specifying the link header. 5.2.3.4 [3] in the spec does not make any references to the POST method and but as it is under the section LDPC/POST it might be seen as it is only applicable to LDPC POST. May be we can refer to 5.2.3.4 from 4.2.4.6 where it says PUT can also be used to create resources. Best Regards, Nandana [1] - http://www.w3.org/TR/ldp/#ldpr-put-precond [2] - http://www.w3.org/2012/ldp/track/issues/91 [3] - http://www.w3.org/TR/ldp/#h5_ldpc-post-createrdf On Fri, Oct 10, 2014 at 10:06 AM, henry.story@bblfish.net < henry.story@bblfish.net> wrote: > How does a client specify that it only wants the creation of a resource, > i.e.. that > it only wants the PUT to succeed if there was not a previously existing > resource there? > Should that not be in the spec? > > Henry
Received on Friday, 10 October 2014 09:31:49 UTC