Mike Linksvayer wrote: > > I prefer the current version. There is no point defining "web > service" here. "web-API" covers that case. >
I agree about that. But I do feel that both "service" and "software service" could have explicit definitions. Neither currently do. (Anyway, how about "Web API" instead of "web-API"?) Also, whilst I remember, there appears to be an aspect that is entirely missing: requiring that all open software service dependencies are also open software services. Dependency services appear not to be mentioned. SUGGESTION """ 3. Whose software service dependencies are all open software services. """ As I said, I'm just looking at how this definition would be consumed by service oriented management. My conclusion: the definition might be better, and the best thing might be to have an open software service definition that anticipates being used in the way I described. One solution could be to have/approve OSS SLAs that conform with the OSSD. Then services which implement an OSS SLA will by definition conform with the OSSD. The OSSD would then remain relatively undefined (for example it doesn't define what "make available" requires). The OSSD need not itself be cast as a service, and could easily remain relatively unapproachable (for example, from the point of view of raising an issue, valid or otherwise). Best wishes, John. _______________________________________________ okfn-discuss mailing list [email protected] http://lists.okfn.org/cgi-bin/mailman/listinfo/okfn-discuss
