Re: Optimizability of FLWR expressions

At 11:04 AM 1/18/2002 +0100, Bas de Bakker wrote:
>Dear WG,
>
>I'm afraid your very procedural description of FLWR expressions,
>together with the short circuiting behavior of the "and" operator, makes
>optimizing FLWR expressions in ways that are now standard for SQL select
>expressions extremely difficult.

Thank-you for the comments and for the detailed examples. These will be 
discussed within the Working Group.

>If you really want to create a well optimizable language, I strongly
>suggest giving implementers a lot more freedom in the order of
>evaluation of expressions.  In the same spirit, I would also like to add
>my support to comment
>http://lists.w3.org/Archives/Public/www-xml-query-comments/2002Jan/0291.
>html
>on lazy evaluation, where similar considerations apply.  Please note
>that in all these cases, the only difference can be between raising or
>not raising an error.  A different evaluation order cannot result in a
>different non-error value.

I agree with you with respect to lazy evaluation. I suspect there is a good 
chance that the Working Groups will be sympathetic to this suggestion, but 
this is, of course, up to the Working Groups.

Jonathan

Received on Friday, 25 January 2002 06:10:53 UTC