Re: Model question

> Date: Mon, 28 Jul 2003 16:40:33 +0200 (MEST)
> From: "Edward C. Zimmermann" <edz@elmyra.bsn.com>
> 
> So do we now start to define search methods and models for the
> individual record types?  some kind of XPath for some XML but where
> its a genetic sequences its sequence into GCG or Blast?
> Or are we not better off.. not sticking out nose into what
> applications of the record the user might
> want.. Yes. Application. Loading a Word file into Word.. or running
> some XPath queries.. or feeding a Postscript file to a typesetter
> etc. are all "uses" of record. Our job I've tended to see as
> providing the mechanisms to enable search and retrivial and to
> create enabling mechanisms for usage of whatever is produced but not
> to go as far as to define what and how the information is to be
> used. If we start to try to define this within the profiles of our
> standards I think we'll be more damned than we've been with our
> already bloated S/R protocol.

I think you're way off-base here, Ed.  What we're talking about is
potential future mechanisms to give users the flexibility they need to
express their own applications' needs.  The ability to treat any given
hunk of data as a database, result-set, record, field or whatever as
the need arises is a way of providing exactly this flexibility.  In
other words, this approach is appealing precisely because it leaves us
"not sticking out nose into what applications of the record the user
might want".

 _/|_	 _______________________________________________________________
/o ) \/  Mike Taylor  <mike@indexdata.com>  http://www.miketaylor.org.uk
)_v__/\  "A program is never less than 90% complete, and never more
	 than 95% complete" -- Terry Baker, IBM's Federal Systems
	 Division, 1975

--
Listen to my wife's new CD of kids' music, _Child's Play_, at
	http://www.pipedreaming.org.uk/childsplay/

Received on Monday, 28 July 2003 11:03:38 UTC