- From: Steve Speicher <sspeiche@gmail.com>
- Date: Wed, 28 Nov 2012 09:17:15 -0500
- To: Thomas Kurz <thomas.kurz@salzburgresearch.at>
- Cc: public-ldp@w3.org
On Wed, Nov 28, 2012 at 9:09 AM, Thomas Kurz <thomas.kurz@salzburgresearch.at> wrote: > Hello everybody! > > I am involved in the Apache Marmotta Incubation proposal > (http://wiki.apache.org/incubator/MarmottaProposal) and have 2 questions > regarding the current working draft of October 25th. > > On 5.2.1 it says a LDP Container is a LDP Resource. > On 4.4.1 it says on HTTP PUT it must replace the persistent state of a LDPR > with the entity representation in the body. > On 5.5.1 it says, servers should not allow HTTP to update LDPC's members. > > This is somehow inconsistent. As 'must' is stronger than 'should not' (in my > opinion), 5.5.1 is redundant. Why is 5.5.1 necessary? Servers get decide when they accept PUT to update/replace a resource. The guidance in this draft is that replacing/updating the membership by using PUT provides a fairly large burden on both the client and server, especially for containers with many members. > I even do not know how an existing LDPR can be added to a container without > PUT (what should be possible somehow regarding 5.2.2). It could be added by POSTing to the container, which creates the LDPR and adds it. It could be added by means outside of what this draft currently says. For example, PATCH, SPARQL Update or something else implementation specific could be used. > Many thanks for your answers and best regards! > Thomas As you may be following, some of this is changing as the FPWG has some open issues regarding some of these topics: PATCH, adding members to existing containers, composition/aggregation use cases supported, etc. Stay tuned and interested in your feedback. Regards, Steve Speicher
Received on Wednesday, 28 November 2012 14:17:43 UTC