I think that it would be a mistake to think of data as some sort of abstraction of a file. Certainly file-like things are an instance of data. But there are others: relational databases, object oriented databases, XML databases, streams of data (although streams are out of scope for V1 of the data architecture). Allen Luniewski IBM Cross Brand Services IBM Silicon Valley Laboratory 555 Bailey Ave. San Jose, CA 95141 408-463-2255 408-930-1844 (mobile) "Donal K. Fellows" <donal.k.fellows@ To manchester.ac.uk> guru prasad <guru_bn@yahoo.com> cc 07/07/2007 10:43 Allen Luniewski/Almaden/IBM@IBMUS, AM Dave Berry <daveb@nesc.ac.uk>, ogsa-d-wg@ogf.org, ogsa-wg <ogsa-wg@ogf.org> Subject Re: [ogsa-wg] [OGSA-D-WG] Definition of "data service" guru prasad wrote:

Thank you for your comments Allen.

In that case I vote for Dave Berry's definition:

"A service that provides interfaces to the capabilities and data of one or more data resources within a service-oriented architecture."

There is a separate question as to whether the data *is* a capability; I suspect it is, at least in some abstract sense. On the other hand, you might wish to distinguish the data as a special kind of capability that needs non-generic handling.

and the Data Resource is .....

A data resource is an abstraction of a file, so I'd imagine that you'd want to describe it as a resource that is an ordered collection of organized information. (The idea is that it doesn't have to be in a file, but it does have many file-like characteristics.) It's strongly beneficial to the overall OGSA if you can have data resources being a subclass of resources (by referencing the OGSA glossary). Donal.