[infod-wg] Minutes from 16 June conference call

Steve Fisher S.M.Fisher at rl.ac.uk
Mon Jun 20 03:10:11 CDT 2005


On Sun, Jun 19, 2005 at 05:40:56AM -0700, Dieter Gawlick wrote:
> Steve,
> 
> I think that the specification has plenty of information about the 
> subscription. I attached a note that describes the various features that 
> we have captured so far; I hope it is not too much a repeat. Obviously, 
> the use case are the place to show cases this specifications and 'bring 
> them to live.'
> 
> I tried to highlight the various functions of subscriptions very 
> carefully; however, the real work has to be done in the use cases.
> 
> Dieter

Dieter,

Thanks for the note. I don't think anything in your note suggests that
"consumptions" cannot usefully be bundled in with subscriptions. I am
not arguing against the properties I just want to grouup them
together.

Steve



> 
> 
> Steve Fisher wrote:
> 
> >On Thu, Jun 16, 2005 at 09:37:23AM -0700, Cecile Madsen wrote:
> > 
> >
> >>
> >>
> >>Attendees: Shailendra, Chris, Dieter, Steve, Stephen, Susan, Vijay, Arjun,
> >>Susan, Cecile
> >>Apologies: Abdeslem
> >>
> >>Minutes:
> >>
> >>We reviewed John's summary/understanding of INFOD specs:
> >>
> >>His review is an accurate summary of INFOD specs; his assumptions are
> >>correct and his comments summarize earlier comments we've had
> >>on the specs (uniqueness of disseminator, unclear how registry knows
> >>whom to notify, simplifying the interface section by concentrating on
> >>artifact
> >>descriptions, not duplicating operations, etc.)
> >>
> >>Discussion on need for subscriber entity; resolution is to describe a
> >>scenario that pinpoints that need, namely attaching vocabulary to
> >>subscriber so that consumers can filter on such information.
> >>
> >>Discussion on weight of subscription; John/the specs don't position
> >>it as the first class artifact that it is. Need to work more on this.
> >>Subscriptions have evolved historically (subject-based, content-based,
> >>state, etc.) and INFOD should position its subscription on top of the
> >>chain...
> >>it can also drive publications...
> >>   
> >>
> >
> >While trying to do my use case (at last) I wondered where to put the
> >information which is used to express the interest of a consumer in
> >some information or potential information. I looked at Cecile's note:
> >
> >INFOD_resource_creation_scenarios-1.doc
> >
> >and noticed that all such information is going into the
> >consumption. There is *no* mention of subscription information.
> >
> >I think that (some of) our problems will be solved by merging the
> >consumption and subscription into something which we just call the
> >subscription. Part of the subscription can be how the data should be
> >processed. I am sure this will be much easier for people to
> >understand.
> >
> >Steve
> >
> >
> >
> > 
> >
> >>TO-DOs by next conf call:
> >>
> >>- Cecile - use John's suggestion - simplify interface section of specs
> >>- Stephen - review the Animation scenario wrt specs - Cecile to help with
> >>questions
> >>- Dieter - follow-up on a to-do from June f2f (send description on
> >>retention)
> >>- All - consumption terminology still open...
> >>- All - review summary notes from f2f on May 31/June 1 for to-dos that 
> >>were
> >>assigned
> >>
> >>
> >>Next Conf Call:
> >>
> >>8AM PST June 23  (note earlier time again... separate note will be sent
> >>with meeting notice)
> >>
> >>
> >>The INFOD team
> >>   
> >>
> >
> > 
> >







More information about the infod-wg mailing list