[tmql-wg] How to proceed with TMQL?
Sun, 14 Mar 2004 16:45:53 +1000
On Wed, Mar 10, 2004 at 10:49:16AM +0100, Steve Pepper wrote:
> I would like to suggest a session (at the start of the meeting?) in
> which each person who has put forward a language proposal gets to
> state his very clear opinion. In other words, Robert, Dmitry, Rani
> and Lars Marius should tell us which alternative (AsTMa? / TMPath /
> toma / tolog / some hybrid) he thinks we should chose as the basis
> for TMQL and why. They should point out very clearly the strengths
> of their chosen alternative and the weaknesses of the others.
1) On the negative side I see that - compared to the RDF universe -
we have too few proposals. Some languages only have tutorials and
no dedicated specification. And we have too few really implemented
proposals. This is a risk factor.
2) On the panic side I read DAWGs schedule:
Having seen their solutions and also the amount of theoretical
underpinning will make us look as amateurs :-)
3) On the positive side, fewer proposals might make it easier to
converge. Or not.
> This probably means they will have to sing the praises of their own
> proposals, which I know will be difficult because they are all so
> modest :-)
I do not think this will be effective enough for what we plan. All of
us can be expected to know about the individual proposals, also in
some detail, given these are documented.
The inverse approach would be that we all prepare statements what we
do not like about the other approaches. At least in my case this has
forced me to question particular things I have done with AsTMa?
Given these, we could walk down the use cases and could try to draft a
new language. I could imagine that a core language emerges so
quicker. At least if we assume we all survive the discussions without
major injuries. I'll bring my body guards anyway...