Agenda

1. Update on core activities
2. Declare AM 2.0 spec final
3. Discuss topics for next release
   • Tom's scenario (change sets, streams, baselines, workspaces)
   • Resource type discovery mechanism for clients
   • Domain ontologies
   • Transactions for highly fragmented resources
   • Control processing (i.e MDD)
   • Core compliance checklist

Attendance

Regrets: Tom Piccoli

Atendees:Clyde Icuspit, Ian Green, John Crouchley, Peter Yee, Sandeep Kohli, Jim Conallen

Minutes

AM 2.0 Specification declared final. no more changes or work will be done on this particular version of the specification. The only remaining activity is to get the RDF vocabulary document published on the open-services.net web site. While the specification will not change, we may elaborate the examples and scenarios if we find that new adopters are having problems understanding it.

The remaining time was spent discussing potential topics for discussion (live or via mail lists).

The list so far:

Model management with change sets/requests - This scenario originated with a scenario that Tom Picolli originally presented to the workgroup that describes a workflow of someone making changes to a model (AM resource) in the context of change sets and requests that can provide a trail of activity and automatic associations to other types of resources.

Resource Type Discovery - As AM service providers offer customer defined resource types it would be useful for clients to be able to find out exactly what types of resources a provider manages, as well as allow the creation of. Presently a client must navigate all the service discovery documents looking at all creation and query services and the resource type and shapes associated with them. It is not clear how a client would know that a provider can manage a given type outside of creation/query.

Meaningful integrations - While simple linking to generic opaque resources provides some benefits, we need to be thinking a little deeper. How do we express meaningful relationships where the providers know about and make use of deeper semantic content through common ontologies and vocabularies.

Transactions and locking - In the AM domain resources are often highly fragmented. When clients are creating or modifying a set of resources enmasse in a concurrent environment it is very difficult to maintain the integrity of the resources. Traditionally we've used transactions and locking to help manage this. Is it appropriate to use mechanisms like this to help control the integrity of the highly fragmented resources.

Control processing, generation and script execution - How do clients initiate processes that create and modify resources in a long running process (ie code generation from models, or simulations).