There are a number of CTS tests in the under review bucket that Tuscany is failing because the assertions tested prior Tuscany behaviour, rather than spec constraints. e.g. SequenceTest.testGetProperty.
These assertions are of the form "get property by index and ensure its the same as getting it by name". and now that Tuscany has stopped exposing EMF special properties the indices don't match up with the property names any more. The hard coding of these indices makes the tests fragile. I could just go round fixing all such instances, however, it seems that we have a lot of occasions where we are testing exactly the same fairly uninteresting code path, and I'm pretty sure that this would be the same for other implementations too. So I propose to rationalize the number of times we do this kind of test, so that we keep the level of redundancy in the tests to a minimum. Kelvin.