Editorial note on resolution logic for error handling

[Juha Tiihonen <juha.tiihonen@aalto.fi>]

 

Assigned to Juha: Resolve Editorial note on line 659-664: FFMII-Specification-Foundation-20111117-v1.0-wd01

5.3 Work Request Status Change Notification

…

Manager-side constraints

Throughout the appropriate Capability Descriptor, a Manager MAY indicate certain constraints in regards to processing of work request Status Change Notification messages. An Implementation MUST comply with any constraints Manager indicates in such case. The full set of Manger-side constraints is described in Section 8.3.3.2.1

Editorial note: Especially in case when Manager is able to accept notifications regarding several WRs in single notification and at the same time in-order-delivery is required, fine-grained resolution logic for error handling is required. The objective is to isolate potentially broken/inconsistent work requests (and handling of notifications thereof) so that they do not affect smooth flow of notifications related to other work requests. This topic is to be in detail discussed in Change Notificaiton Management and Message Handling sections.

 

6.4.3 Message Handling

…

 

 

6.4.3.3 says:  Additional information Detailed requirements for Message Handling layer are further discussed in Section <Link>.

 

Observation: Above link for Detailed requirements for Message Handling layer is broken.

Such section does not exist.

 

Observation: Section 8.1.3.3 ‘Message handling capabilities’ has only a placeholder.

 

Observation: 6.4.3 describes basic Error Indication and Error Handling behavior. However, I was unable to find a section that discusses Message handling and notifications regarding several WRs in single notification. Thus, capabilities to isolate potentially broken/inconsistent work requests (and handling of notifications thereof) were not found.

 

Observation: Batch operations are described (8.7.3). I understand this to be a separate aspect. However, it describes a way to bundle responses.

 

Conclusion: It seems that that the editorial note has not been addressed, and a more detailed section on error resolution logic is missing from the current document.

 

Best regards,

 

Juha