registry roadmap

My feeling is that the structure of Tony's roadmap is correct, but I agree with Bob's instinct that the version reached by this summer MUST be complete and stable enough to allow concrete deliverables by the major projects by the end of the year using the agreed standard. Naming them V1.0 is a good way of signalling this. So basically I think we need to agree a minor re-write, but one that could have a major impact psychologically.

Having said that, I understand Tony's worry that any such standard this year must be preliminary - in a year or two or three, technology will have changed, we will understand the problem that much more, and the bar will be raised with respect to the degree of software automation expected, etc. So it must allowed or even expected that V2.0 could be a radical overhaul, not just a refinement.

Peter and I, as current chair and deputy chair, have just been discussing this, as it is going to be a recurring problem in IVOA. An obvious example is that we are continuing to refine and improve UCDs, while discussing various "fresh start" ideas, such as PCDs, atoms, etc. Perhaps every WG needs two streams - developing current standards, and debating "next generation" standards. Its not clear how we number these streams. V0.8 is a step towards V1.0, and V1.3 is a refinement of that; but is V1.8 a refinement of V1.0, or an early version of a possibly radically different V2.0 ?

I think this should be debated at the plenary in May.

                      andy