This is probably off-topic; I hope someone can tell me where it should
properly be raised.

In a schema I have a Category table related m:m to itself via a Subcategory
table, allowing the recursive construction of taxonomies of indeterminate
depth. There is also a Thing table, and a ThingCategory table that allows
things to be assigned any number of classifications.

It is possible for categories to apply to a thing indirectly by implication
of what might be described as taxonomic inheritance. (In fact the taxonomy
is potentially a network rather than an hierarchy, so there can be multiple
inheritance, but that is another matter.)

I am somewhat at a loss as to how to model this arrangement even using EJB2,
and I am looking for guidance, or at least constructive discussion. Should I
attempt the resolution of applicable categories in a method of a stateless
session bean or with some kind of entity bean?

I doubt it can be expressed declaratively, since most flavours of SQL cannot
tree-walk (recursively resolve self-joins) - this is the classic
explosion-of-parts conundrum endlessly chewed over in the relational world.

Perhaps there is already machinery to handle this situation. If so, I should
be pleased to RTFM if someone would be so kind as to indicate which FM to R.
If, as I suspect, the case is as yet unconsidered, then it would be well to
work a general solution into some future release of EJB; such a thing would,
I suspect, be profoundly useful, allowing straightforward and compact
representation of the recursive classifications so common in the natural
world.

Peter Routtier-Wone
---------------------------------------
One day, all this too shall parse.



--
--------------------------------------------------------------
To subscribe:        [EMAIL PROTECTED]
To unsubscribe:      [EMAIL PROTECTED]
List Help?:          [EMAIL PROTECTED]

Reply via email to