INFOD notes from GGF16 - 14th Feb 2006 -------------------------------------- The lack of Susan's chocolates was noted. Steve Fisher gave the introductory talk Stephen Davey gave a talk prepared with Ronny Fehling explaining the steps to go through to use INFOD. Steve Fisher concluded with our plans. The following items came up in discussion (as noted by Chris Kantarjiev) - ws-eventing should not be under consideration. It is not a public spec. - Where is data sent at the end of the animation job? infod is only sending the notification of status change, not the animation data. - How do we name things, like the information in the wsinfod:RegisterDataVocabulary? Our intent is that it is the EPR of the service that will deal with it, rather than a 'name' that has some implied semantics but might be misspelled. - Should we really allow VocabularyVersion to be an arbitrary string? - Are we starting to develop a language for the Data_Constraint? The data constraint language is not defined, but is tied to the property vocabulary language... - Is there a restriction on the XML schema that is passed in, requiring it to be flat? Not really, since we're not defining the data constraint language in the spec - it's just something we're passing along but is meant to be understood only by the publisher (and the registry, enough to match things up). - Have we defined vocabularies for Schedule/Policies/Processing? No, but we have some ideas about what should be there. - Referring to the presentation of Stephen Davey: are any of these steps optional? You don't need to define properties/propertyValues. But all the rest are needed. Should we be trying to come up with a version that has fewer steps? The descriptions of the vocabularies are complicating things. Some reordering of the material may help. - We should start thinking about the implications of the interop fests that people are discussing, even though our implementation is some 2 years away. - OGSA architecture docs don't mention INFOD!