RE: Ownership of data[bases] (was: SQL Syntax)

> From: David Gibbs
>
> Question related to that ... would you advocate restricting ACCESS to the
> data (in it's raw form)?  What about access to "Logical" data
> (defined sets
> of data with logical data appropriately joined)?

In a perfect world, there would be a completely transparent tier of data,
separate from the actual physical representation on the disk.  This tier
would perform all necessary security, do denormalization as needed, exercise
foreign keys, perform calculations and whatnot.

Raw data access would be completely hidden from the user, and any
manipulation of the data would be through a controlled methodology.  This
would ensure integrity, and also allow the Nirvana of "platform
independence", since in theory the data could be stored anywhere on any
platform, and it would be the middle tier's job to get the data.

Wouldn't that be lovely?

Well, in the prehistoric days, this was known as an "application".  But
since user's needs to get data have outstripped the ability of IT
departments to provide it (for reasons to various to address here), the
trend seems to be towards abdicating that responsibility to the user.
That's a bad trend, in my book.

What's the right answer?  I don't have one yet.  OQL has possibilities, and
so does message-based client/server.  All I know is that ODBC is not the
answer that I would choose.

Joe