Is supertyping orthadox?

Hi all, I am implementing a system that includes a relational database. Simply put it is a system to provision services. Each service has its own independant data structure requirement.

We've had initial design discussions with a firm of consultants and during this came up the issue of how do we store the services data.

The solution they came up with was what they termed as a 'generic table structure' where basically there is a 'schema' table that stores the Field name, data type, length, (also has a foreign key to a 'tables' table to group the schemas together). Also there is a 'values' table that has a foreign key to the 'schema' table that in the implentation would have a sql_variant field that stored the data.

This is fine.. apart from the fact that you are limited to not having *any* relationships between any of the services, which isn't practical for some of the services. .. adding some form of flexible relationship that's easily queryable seems to me like reinventing the wheel of what the dbms does already for us .. not to mention exceptionally complex and probably slow.

My suggestion was that we have a seperate set of tables representing each services data structure specifically defined and we link them all together by using a 'Services' table. ( I think after doing a little research that this is referred to as supertyping. Please correct me if I'm wrong). This was disputed as being a viable alternative.. with no real other argument other than it's 'unorthadox'.

My main question is, is this method unorthadox?.. if it isn't (as I suspect) are there any papers (preferably online) written by respected database theorists that I could read and point them to. I've never actually implemented this method (although I have come up with the same design in a project that hasn't been created yet) so I cannot prove that it works and they have considerably more practical experience than me in this area, so I need some way of adding quantifiable credibility to my argument.

The only other main argument that was raised against it is that it isn't true referential integrity, as you *could* have more than one set of data in the services structures for the same 'Services' table identifier.. which is true but this would be constrained at business logic level (trigger etc) and would only have performance impact on inserts.

Thank you for your time reading this and for any comments.

Kindest regards
 -=- Brett Received on Tue Mar 20 2001 - 13:28:04 CET