Re: Property sheet, ad hoc, property page, flexible data

"AC" <AC_at_No.spam> wrote in message news:TkFEe.4555$6M3.3607_at_trnddc03...

PMFJI. The fundamental problem with user defined attributes is not a theoretical one. It's one of conflicts between two incompatible requirements of the users.

The first requirement is, as you outlined, that users want to be able to extend the schema to include data items that are not defined in the schema, and are defined in each user's head. Each user's extensions might be understood by the entire user community, but more commonly, each user's extensions are idiosyncratic to that user, and either meaningless or misstated in the view of the rest of the users.

The other requirement is that the data in the data base be sharable and usable by the entire user community in search of information. You can't turn data into information without the semantics. Of all the summaries of the problems of using data not included in the schema, I've never seen a more elegant summary than Marshall's:

No schema, no semantics.
 Know Schema, Know Semantics.

Now, here's the kicker: there's generally a delay time between the time the users demand the ability to add to the schema, and the time the same user community demands that the architects or administrators make sense out of the whole mess. By the time the mess has been made, and the semantics are encapsulated inside the skulls of hundreds of users, many of whom are former users, it's the devil's own work to analyze all that data, and to obtain the kind of outputs that "ought to be easy, because the data is in a database".

If you can somehow get the users to agree that they are responsible for sharing the semantics of the data with each other, and that the data architecture can't help them , except for data inside the schema, then, and only then, many of the suggestions already posted will do the job just fine Received on Sun Jul 24 2005 - 15:10:04 CEST