Objection to "optional match" and "result limit" requirements

Network Inference has objected to both the "optional match" and "result
limit" requirements adopted by the working group at the Carlsbad
face-to-face.

One major concern throughout the standardization process is that the
group could easily be drawn into new research, and end up standardizing
a language based on a formal model which is not fully understood by the
community, or even the group adopting it.

"Optional match", in particular, worries me in that it necessarily
complicates the model underlying the query language. Beyond purely
theoretical concerns about the ramifications of this addition, a more
complex model will almost certainly make any query language more
difficult for users to understand.

"Limit" is also clearly outside the simple algebra of variable bindings
and triple patterns which is described by the other adopted
requirements. With no definition of ordering, a "limit" feature violates
many properties which can be desirable for a query language, including a
simple desire for deterministic results.

In both cases, Network Inference certainly sees value in both features,
and supports both as objectives for this working group. If the potential
problems related to these requirements can be overcome, then our
objection to the classification of these features as "requirements"
should not prevent the group from regaining consensus on a final
recommendation. However, Network Inference encourages the group to
remain open to the possibility of a recommendation which does not
include these features in its initial release if their addition proves
to be problematic.

Received on Sunday, 18 July 2004 20:59:57 UTC