I had a phone call yesterday with Nicolas Rouquette, Hans-Peter de Koening, and Roger Burkhart of the QUDV group. Summary:
* They will consider removing "Quantity" from their model because it is really what we call "quantity value" and they actually model quantity values as what they call <<value type>>. A <<value type>> is a number (which may be real, integer, etc.) with a unit of measure and a quantity kind. So the main difference with our "quantity value" is that our version does not include "quantity kind".
* There is a usage difference between <<value type>> and our "quantity value". In SysML, the unit of measure and/or quantity kind may be "precommitted" when defining a kind of <<value type>>. For example, one can define "PencilLength" as a <<value type>> that uses centimetres. In SBVR, a domain vocabulary could use a Necessity for this purpose. And our "duration" and "date time" concepts are already limited to the "duration" quantity kind.
* QUDV has no support for compound quantity values. They suggest looking at MARTE for precedence in this area -- but I think this is overkill since a compound value is just a combination of atomic values.
* We discussed our use of "categorization type" (UML powertype) for "quantity kind". SysML avoids using the more complex aspects of UML, so isn't in a position to do this. We did not reach any real resolution on this difference between Date-Time and QUDV. We may just have to live with this.
We will have another phone call, tentatively on January 11. Anybody from the date-time group that wants to be involved, please let me know.
--------------------------------
Mark H. Linehan
STSM, Model Driven Business Transformation
IBM Research
phone: (914) 784-7002 or IBM tieline 863-7002
internet: [email protected]