Re: Thoughts on the LDS WG chartering discussion

> On 10 Jun 2021, at 13:55, Peter F. Patel-Schneider <pfpschneider@gmail.com> wrote:
> 
> On 6/10/21 6:28 AM, Ivan Herman wrote:
> 
>> Graphs/Datasets are more often than not store[d] in datastores, triple stores, knowledge graphs, you name it. 
> Agreed.
>> The serialization format used to feed the triple store is irrelevant, 
> Agreed..
>> and clients of such triple stores may request the data in different serialization format that suits their needs.
> Agreed.
>> If the consistency of such graphs (ie, set of triples or quads in the triple store) has to be checked
>> via, say, a hash, then the approach you are describing does not work, due to the problem of bnode labels: triplestores are free to relabel the bnodes of incoming graphs and producing new labels when they export them.
> 
> How does "consistency" fit into this?  Every RDF graph (or datastore) is consistent.

I am sorry, wrong choice of words. If you want to check that the graph you retrieve from the data store has not been tampered with; e.g., by checking its hash.

An analogy is a number of open source sites where one can download an application and check the hash value of the downloaded package against the hash of the application announced somewhere.

Ivan

> 
> If you mean that one wants to determine whether one RDF graph (or datastore) is isomorphic to another, then just serialize the contents of one in N-Triples (or N-Quads) and verifiably send that document to the other where the document can be deserialized and the result compared  for isomorphism with the local graph (or datastore).   If you want to optimize the process then, yes, you can canonicalize each graph independently into N-Triples (or N-Quads) and compare hashes of the documents but this is not necessary.
> 
>> 
>> Also: isomorphic graphs do not have the same hash value, because graphs may be ony be isomorphic via a suitable relabeling of bnodes.
> Graphs are isomorphic or not.  Relabelling of bnodes does not affect isomorphism.  This is necessarily the case because bnode labels are not part of the RDF abstract syntax (data model).  In any case, an RDF graph isn't a document and doesn't directly have a hash value.
>> 
>> Ivan
>> 
> peter
> 
> 


----
Ivan Herman, W3C 
Home: http://www.w3.org/People/Ivan/
mobile: +33 6 52 46 00 43
ORCID ID: https://orcid.org/0000-0003-0782-2704