Dr Robert Sanderson wrote:

> On Mon, 20 Dec 2004, Adam Dickmeiss wrote:
>> Dr Robert Sanderson wrote:
Just some short questions to a long debate, which seem to wander off the core problem - if any such core problem exists in practice.

Does anybody had problems with SRU/GET yet?
Problems which will corrected using SRU/POST?

I mean, in practice, not in theory ??

How many of the existing implementations do hit this SRU/GET problem ??
How many of them will coose to solve it by a third SRU/POST protocol ??

Unless there is a proven problem with SRU/GET which can be corrected with SRU/POST, and sufficient many people do hit the problem, in my humble opinion, there is not much need to add a third protocol type to the specs.

I fear that you are out to invent standards faster than the are in use out there. IMHO it is better to relax a little and see how the SRU/GET
gets adopted to the search market, than making the standard complicated, and make more technical hurdles for fast adoption. Standards should standadize existing practices, but not drive implementations into many different implementations.

This said, I think it is valuable to debate the issue, and to make test implementations of SRU/POST to check out how they will perform, and what problems they will solve. I will though hesitate to impend this on the standard before usefullness is proven in practice.

Cheers, Marc Cromme, Index Data