Not being a former PTC employee, and knowing this was technically acquired software that has essentially been rewritten, I can only share my technical perspective:
Rational Rose from IBM is used to model the java object to relational mapping in schema. They are all primary and foreign keys. 95% of the time there is a cooresponding column ending with same characters classnamea2a2 etc.
You know to join the two tables by these keys.
Back to modeling, all link classes that provide more efficent lookup of one-to-many or many-to-many relationships typically extend the ObjectToObjectLink class. It has in it model wise a roleAObject and a roleBObject. This simple fact generates well over half of the column names for IDA3B4 and IDA3A4,etc.
You'll always notice the primary key is IDA2A2. I can only guess that IDA stands for IBM InfoSphere Data Architect and 2A2 is the hexadecimal identifier.
Clues are always going to be found in the model ( high abstract levels like Persistable and WTObject).
What I can tell you is that as you add additional columns to a modeled object that alias other modeled objects, the number used increments: for 3 references, you might find and IDA3A3, IDA3A4, and IDA3A5 in same table and same columns names used on different tables for different purposes.
My honest guess is the entire notion of the columns named the way they are have to do with the tools used to generate the object model, and not anything deliberately done by PTC.
I could be wrong and surprised, but I doubt you'll find anyone that will give you 'the answer' as many a folk have moved on from PTC in early days.
Now that Rational Rose is not required and you may use Java Annotations starting in WC 10, could these change? Not sure, the clue also could be in the wt.pds package and it handles conversion to SQL for object relational mapping.
Regards,
David DeMay
Sent from my Verizon Wireless BlackBerry
