Web Services Data Access and Integration - The Core (WS-DAI) Specification https://forge.gridforum.org/tracker/index.php?aid=1711 Thanks to those who have provided comments on this and the descendent specs. Here are some notes on how we propose to respond. #1 Comments_WS-DAI_from_Oleg.doc https://forge.gridforum.org/tracker/download.php?group_id=90&atid=414&file_id=736&aid=1711 Response: Oleg has provided a collection of comments and suggestions relating to the clarity of the document and its consistency with terminology used elsewhere. We propose to go through these carefully during the revision process, making changes either as proposed, or at least to try to address the questions raised. We do not envisage any changes to the substance of the specifications from these comments, and thus do not discuss the individual comments here. #2 Comments WS-DAI from Dave Berry.doc https://forge.gridforum.org/tracker/download.php?group_id=90&atid=414&file_id=752&aid=1711 Response: Dave has provided a collection of comments that, like those from Oleg, relate to the clarity of the document and its consistency with terminology used elsewhere. As before, we propose to go through these carefully during the revision process, making changes either as proposed or at least to try to address the questions raised. Several of the comments made merit a more focused response: Comment: Has this document been reviewed by the OGSA WG on an OGSA telcon or at an OGSA F2F? I know DAIS has had good links with OGSA; I think it would be beneficial to both groups to have a review of the final document. Response: There hasn't been a formal group-to-group interaction. I posted to the OGSA-WG to let it be known that the DAIS specifications were entering public comment. As he is Chair of the OGSA Data Architecture Working Group, my inclination is to see the comments from Dave as representing informed feedback from the OGSA community. I think there have been quite a lot of chances for OGSA-WG people to comment (e.g. through public comment and at GGF16), and am not sure that I think we should explore further mechanisms. Of course, we will be happy to contribute to OGSA Data Architecture WG Sessions at GGF17 if this is felt to be appropriate, but there has been quite close engagement over several years. Comment: How should users find what other URIs are supported? I suggest that GGF maintain a registry of data formats, which vendors can update as appropriate. There’s no need to record every XML schema, but every description language, of which XSD is one, should be recorded. Response: I agree that this could be useful. Outside the scope of the DAIS Working Group, though. Comment: I suggest that DAIS defines an initial set of language URIs, particularly for established languages. Then GGF should maintain a registry of language URIs that vendors etc. may update with their own URIs. Response: This seems plausible for the XML query languages, and we will check with XML database specialists that it makes sense to do this in WS-DAIX. Alas, it's a known open issue for relational databases that we don't know how to solve. Comment: A free format description of the data resource represented by a data service. I don’t understand the intention - is this to allow a human-readable description? Response: Yes. #3 Michael J Beckerle Issue: The issue is with the parent-sensitive to child and vice versa thing. I expect the most common form of sensitivity is the "unknown" form i.e., where updates going on might or might not be visible within a row-set or by way of an indirect data service. The current enums let you say that you will see them, or you won't see them, but not "you might see them". Comment: I agree that the current terms state either that you do or you don't, and don't accommodate uncertainty. Options include to extend the vocabulary as suggested, or to revise the existing definitions to indicate that it is "insensitive" if no change will ever be reflected in the related data resource, and that it is "sensitive" otherwise. As such, no guarantee is provided that a change will be reflected in the related data resource, but an implementation either includes functionality to reflect the change or not - if it includes and activates such functionality, then it is sensitive. This needs discussed around interested parties to reach a consensus. Please can people with a view post to the DAIS-WG list.