I've been following with interest the discussion of next steps for OGSA-WG, and also talking to various people in the community. Based on this, I'd like to express my views that:

a) The single most critical thing we can do as we develop OGSA v2 is to obtain broad critical input and buy-in from informed and influential members of the IT community. We have been successful in establishing OGSA as a "lighthouse" that people look to as a source of direction for their efforts. Needless to say, the lighthouse is a bit rickety and dim in places, and not everyone is steering by it. So we have much more to do.

Thus, I'd like to see us talking at length at the F2F about what sort of input we need, how we get it, and how we integrate it. E.g., we might want to see our participants from HP, IBM, Sun, Intel, etc., volunteering to take what we have to the appropriate internal product groups to get deep critical review and feedback.

b) There is continued interest being expressed in "going to the next level" in terms of level of detail. I think this can be useful if strongly motivated by external input, but otherwise is very dangerous, as our constituency of IT vendors and users, while enthusiastic about a document that lays out a vision of what needs to be done, will I think be far less enthusiastic about a document that tells them how to do it, if they have not been part of the process.

Thus, I'd like to see discussion at the F2F about where to draw the line. I see the design teams going periously close to crossing the line into detailed specification, and I believe strongly that this work is not appropriate for OGSA-WG. I know we've discussed this before, but I think we need to revisit it again.

Here's one approach that I'd like to discuss in San Jose: Let's assume that the current level of detail in the OGSA-WG specification is close to adequate, although not uniformly so. Let's say that anyone going further has to start a WG to do so. And furthermore, as OGSA aspires to be an official GGF architecture, let's set the bar high for products of such WGs to be part of OGSA: e.g., let's follow GGF's processes for making something a standard--i.e.,multiple interoperable implementations and substantial adoption. That way, we don't integrate any true specifications into OGSA until they have passed the tests of implementability, usability, interoperability, and acceptability.

I look forward to seeing many of you in San Jose.

Regards -- Ian.