Re: the state of ldp-patch, and a procedural proposal

John Arwe <johnarwe@us.ibm.com> wrote:
>For the portfolio of implementations I ride herd over, blank nodes are 
>very rarely used because we expect over time to need to index much of
>it 
>and that's just easier if there's a URI (even a hash URI

Just out of curiosity, what exactly do you mean by hash URI?

, which is how 
>most of our implementations get rid of their blank nodes).  It's
>usually 
>easy enough to generate a hash URI algorithmically, hence consistently,
>
>for the kinds of data we're exposing today.
>There are a few places that have RDF lists etc, so they're still
>lurking, 
>but we're careful to allow use of those constructs only in parts of the
>
>model that we don't expect to query over later with any regularity. 
>And 
>more central to this discussion is that blank nodes are typically used
>in 
>what LDP calls "server managed properties", so there is essentially
>zero 
>chance of them processing PATCH requests over them.
>

So you'd be happy with a PATCH that had support for list operations but not blank nodes?

Do you ever have a subject s, a predicate p, and TWO values, v1 and v2, BOTH of which are lists?

If you were ever to have that, can you think of any way you could indicate in a patch which list was to be modified?

    - Sandro

>Best Regards, John
>
>Voice US 845-435-9470  BluePages
>Tivoli OSLC Lead - Show me the Scenario

-- 
Sent from my Android phone with K-9 Mail. Please excuse my brevity.

Received on Wednesday, 2 October 2013 20:44:00 UTC