> Date: Tue, 2 Sep 2003 14:43:29 +0100
> From: Robert Sanderson <[log in to unmask]>
>
> [boundingRect(13N25W,25S70E)]   is a valid cql term without ""s...

Nope.

> (as there are no spaces)

But the open paren is a special character ... which is why it's
posible to parse queries like:

        foo and(bar or baz )

> I like the relation modifier approach, personally.  When a relation
> structural modifier is defined, then there obviously needs to be a
> structure along with it.
>
>   time within/IsoDate "2003-10-10 2003-11-20"
>
> rather than
>
>   time within "[isoDate(2003-10-10)] [isoDate(2003-11-20)]"

Yes, this is a compelling line of reasoning.

 _/|_    _______________________________________________________________
/o ) \/  Mike Taylor  <[log in to unmask]>  http://www.miketaylor.org.uk
)_v__/\  "If you really want to be a writer, develop calluses on your
         pride" -- Adrian Bedford.

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