Multiple stereotypes in EA data model

[Paolo F Cantoni]

Those of you who have dabbled in the EA data model within the EA repository will know that stereotype information is held in two places:
The primary stereotype is held within the appropriate table for the type of item.  the auxiliary stereotypes are held in the t_xref table with an GUID link back to the appropriate item in the main table.

Now, we have some repositories from a few years ago that we are trying to "bring up to date".  It won't surprise you to find that we have found inconsistencies in how EA appears to have handled stereotypes over the years.  

We want to standardize the stereotype behaviour so that it matches what EA would now produce (v7.1 build 832).

It would seem that EA's behaviour NOW is:

1) When it creates the primary stereotype, it also puts a matching entry in t_xref.

2) Alternate stereotypes are held only in t_xref

3) Apart from the primary stereotype, the order of the auxiliaries in t_xref determines the order of the auxiliary stereotypes on the UI.

Can anyone (preferably from Sparx) confirm or deny this analysis?

TIA,
Paolo