RE: datatyping is not needed

OK, I finally caught up on reading all my WebDAV e-mail. Sorry it 
took me so long. So, now I can respond properly.

> Well, I meant to make a claim more analogous to this one:
>  1) Some drivers want to travel at 100 mph
>  2) DASL has to choose between walking and a bicycle
>  --> DASL should walk, because those who desire speed will 
> buy a car, and wouldn't use the bicycle anyway.

I offer the following (hopefully entertaining) fable in good 
spirit and with all due respect to illustrate my point.

Once upon a time, a traveler wanting to get from 
point A to point B encountered a professor. "I want 
to get from point A to point B in a hurry", the traveler said.

"Well, this is point A. Point B is ten miles thataway. 
So you walk in that direction," said the professor.

Whap! The traveler was in no mood to suffer fools
gladly, so he slapped the professor up side the
head to get his attention. "I want to get from point
A to point B in a hurry," the traveler said.

"Some people want to go 100 mph, and, as you can
see, I have no car, only this bicycle. Logic
dictates therefore, that you should walk ...,"
said the professor.

WHAP!! This time the traveler slapped the professor
up side the head with a textbook on mathematical logic.
"I want to get from point A to point B in a hurry,"
the traveler said.

"OK, OK. I get the hint. My logic was in error, 
right? Here, take this bicycle," said the professor.

The traveler got on the bicycle, and pedaled off to
point B, getting there 10 times faster than he
would have if he had walked. The professor was
happy because he learned something, and the
traveler was happy, because he got there faster.
And they all lived happily ever after. The End. :-)

My point is that we will be justly criticized for
(a) not doing as good a job as we could have, and
(b) having an error in our reasoning,
if we don't require collections to advertise at 
least the names and datatypes of their properties 
(for the properties that have the most basic datatypes 
in the known universe -- number, string, Boolean, datetime).

I see NO DOWN SIDE in advertising information
the query engine MUST already have in order to
do its job. (The server can not compare "10" against
"2" for greater-than, or even "0" against "00" for
equality if it doesn't know whether the datatype
of the property being compared is a number or a string.)

I hope we can drop this entertaining thread now, whether we 
agree on the validity of the reasoning by analogy or not. What's 
important is what the DASL where condition and query schema 
contain. I will send another e-mail reply on what compromise we may 
be able to agree on. I think it will be close or identical to one 
of your suggestions.

Alan Babich

Received on Thursday, 16 July 1998 17:46:53 UTC