Re: protocol draft updated, open issue proposals (fromUnionQuery)

On Mon, May 16, 2005 at 12:55:34PM -0500, Dan Connolly wrote:

> I think the design is stabilizing here, but
> I'd sure like to see a tad more implementation experience with this
> before I put the question.

Ah, actually, in newfound zeal to be done, I confused this *issue* with an
open ACTION that required the addition of HTTP traces to the draft. Having
added them, I wanted to claim victory on the ACTION and confused it with
this issue. Mea minima culpa.

However, yr comments are in line with mine -- about the need to flesh out
the FROM/FROM NAMED and ambiguous dataset issues -- so I'm happy to second
them (or to be seconded by them, whatever).

> I've run queries using just the background graph
> (using either the FROM keyword or default-graph-uri) and I'm satisfied
> that we're on the same page there.

I'd like a trace for the spec that uses FROM. I could write traces by
hand, but maybe it's better if they come from actual network interactions?
I'm pretty blaise about this difference, though.

> Does anybody have code that can run maybe a handful of examples
> using FROM NAMED?
>  - using FROM and FROM NAMED in the QL, with 2 or 3 of each
>  - using named-graph-uri in the QL
>  - using both, showing that the override works and makes sense

This especially important, IMO.

> For those cases, I don't see quite enough evidence that the
> design is mature yet.

Fair enough.

Kendall Clark

Received on Monday, 16 May 2005 19:34:31 UTC