SpecGL observation from talking to XQuery

I will not be able to attend the telecon today. I thought I should
pass along one item from my discussion last week with the XQuery
editors. We talked about more than just the Functions & Operators spec
that I reviewed, because that review was supposed to be a signal of
the spec concerns that will aply to more of their specs.

An implementation of an XQuery processor will be allowed to vary by
being either a static analyzer or dynamic. We discussed how to
characterize the difference. At first it seemed like a profile, since
the variance can be described as whether the user gets errors
reported early (static) or late (dynamic). As the discussion
progressed, we started thinking of it as a matter of levels. Every
query that works in a static processor will also work in a dynamic
one, but the reverse is not the case.

For now, they will think about the variance as a two-level case. With
the collapsing of the module/profile/level guidelines into one, their
initial discomfort (the variance didn't seem like any of those) may
be smoothed over. I don't think QAWG is required to take action, just
note the discomfort in case that verbiage is adjusted again.
.................David Marston

Received on Monday, 4 August 2003 09:44:14 UTC