RE: datatyping is not needed

Yes, you have stated my position correctly.

Having the datatypes in the advertisement of
query capabilities is enough for me.
Clients can choose to ignore it or use 
it. (In fact, clients don't even have to perform
query capability discovery if they don't want to.)
Having datatypes in the "where" condition as well 
would be redundant, and I do not propose we
do that. (I seem to remember that you astutely 
pointed out the redundancy at the Redmond meeting
and that I agreed with you.)

Hopefully we're violently agreeing: Jim says "its
less filling" (the query condition is not decorated
with datatypes), and I say "it tastes great" (the 
datatype information is available in the query
capability information), and we're both right.

But I would quibble a little bit with you on your wording of
my position on the UI issue. I do not claim we have to make 
the UI EASIER. I claim we have to make QUALITY UI's POSSIBLE. 
That's a very different thing. You're right, though, we're
doing a protocol design, not a UI design. A good protocol
design provides all the necessary information, and it
is not unnecessarily hard to generate or process. Right?
Only the ease of generating and processing the protocol
might make the UI code easier or harder, not the
information providable via the protocol.
The UI's potential for quality could only be affected by 
what information is providable via the protocol, not
by the ease of generating or processing the protocol.

Alan Babich

-----Original Message-----
From: Saveen Reddy (Exchange) [mailto:saveenr@Exchange.Microsoft.com]
Sent: July 07, 1998 1:30 PM
To: www-webdav-dasl@w3.org
Subject: RE: datatyping is not needed



I'll respond first to the UI issue. Alan, making UI easier, while a good
thing, isn't a fundamental scenario we have to address. If dropping
datatyping meant that it didn't make this easier, I could still sleep at
night.

As for the second part ... it doesn't seem like you guys disagree (and
you even point this out). Jim's position (please Jim clarify if I am
missing it) is that datatyping is not absolutely needed in the query
expression. Alan, your position seems to be that datatyping is needed in
query schema discovery. Do I have this correct?

Thanks,
Saveen

Received on Sunday, 12 July 1998 23:35:26 UTC