Re: DELETE and INSERT redundant vs. DELETE/INSERT?

Holger,

Thank you for your comments, and apologies for not replying sooner. I
have responded below.

On Tue, Jan 4, 2011 at 9:54 PM, Holger Knublauch <holger@topquadrant.com> wrote:
> I am working on an update to the SPIN RDF syntax to support the latest
> SPARQL 1.1 UPDATE draft. The latter contains a comprehensive
> description of DELETE/INSERT [1], followed by an equally comprehensive
> coverage of DELETE [2] and INSERT [3]. The latter two appear very much
> redundant, and are IMHO even confusing. DELETE and INSERT appear to
> be a true subsets of DELETE/INSERT, and even the Grammar [4] does not
> distinguish them.

These operations are indeed degenerate cases of DELETE/INSERT. They
were initially proposed as separate cases, to avoid a case in the
grammar where both DELETE and INSERT are omitted, but they have now
evolved to be subsets of the more general DELETE/INSERT operation.
They have been left in the documentation as informative clarifications
of very common operations. In this respect, they are directed more to
users than to implementors of the specification, and we see some value
in keeping them, but we have moved them to subsections of
DELETE/INSERT.

> Could those two simply be dropped, to simplify the documents and
> reduce the (apparent) implementation burden?

Rather than removing these sections, they have been explicitly marked
as "Informative" subsection with an explanation that a correct
implementation of DELETE/INSERT will correctly cover these operations.
Furthermore, The new section "4 SPARQL Update Formal Model" in the
SPARQL 1.1 Update document (see
http://www.w3.org/2009/sparql/docs/update-1.1/#formalModel for the
current Editor's draft) shall clarify this behaviour.

Please indicate if this response adequately addresses your concerns.

Regards,
Paul Gearon
(On behalf of the SPARQL Working Group)

Received on Wednesday, 4 May 2011 00:17:56 UTC