[oslc-rm] Issue on redundant inverse predicates

[Steve K Speicher]

Throughout the development of specs within the RM WG and various other 
OSLC domain WGs we've been in discussion on ways to best handle links, for 
example modeling them from an outbound perspective of the current 
resource. 
Time and experience has shown that this introduces statements which are 
redundant, such as 
(ChangeRequest1, oslc_cm:implementsRequirement, Requirement1)
which is exactly the same as saying:
(Requirement1, oslc_rm:implementedBy, ChangeRequest1).

This causes various problems around data duplication and ease of writing 
queries.  This motivation and best practices has been captured in Core WG 
Link Guidance [1].

For the full list of RM (and other) predicates that are considered 
redundant and the recommended predicates, see the full listing at [2].

For ease, here is the list of RM predicates proposed for consideration of 
deprecation:
trackedBy
implementedBy
affectedBy
validatedBy
elaborates
specifies
satisfies
decomposes
constrains

[1] - http://open-services.net/wiki/core/Links/
[2] - http://open-services.net/wiki/core/V2DuplicatePredicates/

Thanks,
Steve Speicher
IBM Rational Software
OSLC - Lifecycle integration inspired by the web -> 
http://open-services.net