A few more notes on section 4:
1) 1302 4.1.1 The Model for Service Description
1303 Figure 14 shows Service Description as a subclass of the general Description class,
- the Figure 14 is absent in given version.
2) in my previous messages I mentioned issues with Contract and Resources in the Service Description. This specifically relates to 184.108.40.206.1 Description Subject and referred 'resource'. Consumer does not care about service resources, it cares about service's Business Capabilities; resources is the private matter of the service provider.
3) I have an impression that sections 220.127.116.11.1 Description Subjec, 18.104.22.168.2 Provenanc and 22.214.171.124.3 Keywords and Classification Terms mess service with resource, which contradicts our previous definitions. These sections better be re=worded or removed.
4) we have definition for RWE, private and shared + we have definition for:
Service Level Real World Effect
1481 A service level real world effect is a specific change in shared state or
1482 information returned as a result of interacting with a service.
1483 Action Level Real World Effect
1484 An action level real world effect is a specific change in shared state or
1485 information returned as a result of performing a specific action against a service.
The first question is why we need such complex classification of RWE (the simple one is RWE=service result)? The second question is: why Service Level Real World Effect is always shared RWE; a consumer may not have its private return from the service or what? The same relates to the Action Level Real World Effect.
I do not understand Action Level Real World Effect. Web a consumer calls/invokes a service a specific action against a service is executed always; we do not have any other model of service invocation. That is, this definition is tautology. The return or result of the service execution (which I take as The RWE) is always a response to the specific action against a service. So, why we have Action Level Real World Effect?
I would understand if we say that the RWE may comprise outcomes from any part of the service during its execution, not at the very end. These partial outcomes are RWEs also. However, in this case the definition of Action Level Real World Effect has to say something like: "...returned as a result of performing a specific service action"
5) I disagree with the style of expression given in the paragraph:
Figure 22 combines the models in the subsections of Section 4.1.1 to concisely relate
1585 action and the relevant components of Service Description. The purpose of Figure 22 is
1586 to demonstrate that the components of service description go beyond arbitrary
1587 documentation and form the critical set of information needed to define the what and
1588 how of action
It is not a style of standard saying "the relevant components", we need to say what particular components are. Moreover, the statement "form the critical set of information needed to define the what and
1588 how of action" in nonsense to me: Service Description (if the service provider is not a suicide-professional) will never expose its internal actions in the service announcement; also, this is going against the SO Principles. Yes, the Service Description can " go beyond arbitrary 1587 documentation" but this should be about Business Functionality , not about internal service's activities. That is, the Figure 22 demonstrates internal, publicly invisible relationships between public Service Description element and internal service's actions.
6) The Figure 30 Interaction dependencies and related text misses Execution Context as an interaction dependency.
7) in several of my previous posts, I have criticized following sections: 126.96.36.199 Service-Oriented Business Processes (tautology); 188.8.131.52 Service-Oriented Business Collaborations (super-tautology) including inaccurate reference to the choreography (which is wrong because business collaboration perfectly goes via orchestration as well. More accurately, it goes via combination of orchestration AND choreography) and doubled definition of 'choreography'.
Also , Figure 35 Abstract example of choreography of service-oriented business collaboration shows business absurd - coupling of internal business process. Which Business School teaches this type of collaboration? To my knowledge, it is possible only after merge or acquisition, but it is a different business case all together.
8) The problem with:
4.4.2 Policy and Contract Enforcement
2603 The enforcement of policy constraints has to address two core problems: how to
2604 enforce the atomic policy constraints, and how to enforce combinations of policy
2605 constraints. In addition, it is necessary to address the resolution of policy conflicts.
- the title includes "Contract Enforcement", which is never mentioned in the section and, IMO, we agreed that Contract differs from Policy in that the Contract cannot be enforced, didn't we?
Guys, I am sorry that I found so many issues but I believe it is better later than never. Better if we fix them than we will be finger-pointed by other later. Standard reputation is 75% of its life.
Sent: Tue, Dec 21, 2010 11:23 pm
Subject: [soa-rm-ra] Groups - New Editors' Draft, 21 Dec 2010 - Clean (Editors' Draft-2010-12-21 (Clean).pdf) uploaded