AssociationReference files (full, delta and snapshot)

ISRS-132

DK do not keep records of the number of components they have updated.

According to the traceability service, between December and March (their editing cycle for this release) they changed approx 350 concepts - HOWEVER in almost every concept they inactivated the old Description and created a new Description, so multiple changes per concept.

The good news is that this matches the diff report, which shows 349 active record changes (i.e.) new description records, and 346 inactivation records.

There are less changes to the Language records than to the Description records, due to the back end fixes that were required as part of the editing cycle.

The description-DA file contains 4 unchanged apart from the effective time due to these fixes.

The language refsets for the two extra inactive descriptions are already inactive, and thus not appearing in the current language refset delta:

id effectivetime active moduleid refsetid referencedcomponentid acceptabilityid
c8a8c6f4-6dc1-4552-9a53-a19b519ca5f0 20110812 0 554471000005108 554461000005103 1999851000005116 900000000000548007
2a82a32f-eec2-4919-bd4d-10f54a6e060d 20110812 0 554471000005108 554461000005103 3942691000005118 900000000000548007

ISRS-134

No new stated relationships were changed as part of the DK editing cycle, and no concepts were updated. These changes are therefore unexpected.

Looking into them they appear to be partially a result of role group churn (due to the known issue in the classifier wrapper), but I've also found several which are new to the March release, and also some that were in the November release as inactive yet have now been activated (with all identical fields including the Role group) in the March release. Some have even been born inactive in this release...

Further investigation therefore in progress...

We discovered that the differences are due to the following:

1. Several of the changes are due to the role group churn issue
2. Several of the changes are due to inactivations in the International edition
3. The remainder of the changes are due to changes in the dependent International Edition which have impacted the Danish extension (despite there having been no changes in this DK editing cycle to either concepts or stated relationships) due to previous modelling in the Danish extension which incurs impact when classifying.

Confirmed on the re-test that classifier wrapper fixes have removed 16 of these differences as expected - remaining differences are due to the reasons above and so are valid differences.