Pending issues for update

Hi all,

Axel asked me to post the remaining issues for Update.

The top issues from my perspective are:

- DELETE syntax is ambiguous. should statements be separable with
optional semicolons? Or is another syntax option better? This was
discussed, but I wasn't clear on the outcome.

- ISSUE-20. Are there differences between an empty graph and a
non-existent graph? I like to think so, but didn't see a consensus.
This issue affects the need for CREATE and DROP graph operations, and
controls whether or not INSERT/DELETE control graph existence.

- Are blank nodes permissible in the template for DELETE? There was
some discussion, but if there was a resolution then I missed it.

- ISSUE-51. Shall dataset clauses be allowed in SPARQL/Update? This is
one Lee raised, so he should have good comments on it. This is also
one of those areas where new syntax has been proposed that I think
makes sense.


The remaining issues:

- We need the grammar defined. That depends heavily on the above issues.

- ISSUE-28. What happens if entailed triples are DELETEd? I presume
this is equivalent to an no-op, but there is no resolution recorded.

- ISSUE-37. How does basic federated query interact with
SPARQL/Update? Personally I don't see the problem, but the issue is
still listed as open.

- We need a definition of SPARQL Update requests.

- ISSUE-26. How far do we go with transactions/atomicity?

- ISSUE-18. Are there concurrency issues to be addressed?

- Need a mention of what happens when a document is only partially loaded.

- Need to discuss the implications (or limits) of the CLEAR command on
the default graph on services that form the default from a union of
other graphs.

- ISSUE-19. What are the security issues around SPARQL Update?


Regards,
Paul Gearon

Received on Tuesday, 16 February 2010 15:05:50 UTC