Re: Why not reinvent the wheel?

[Vasileios Papadimos <vpapad@cse.ogi.edu>]

On Mon, Feb 26, 2001 at 06:06:48PM -0800, Evan Lenz wrote:

> What you claim to identify as a distinct property of XQuery is common to
> XSLT. Both XQuery and XSLT are declarative, not procedural. 

I only mentioned 'declarative' to contrast the relational model
to the 'procedural' flavor of the hierarchical or networked data models.
Both XQuery and XSLT are declarative in that respect.

> Neither enforce any particular implementation strategy. 

I am not sure about that (and this is true for XQuery too!) 
One way of enforcing implementation strategies is overspecifying. 
As an example, saying that joins respect the ordering of their inputs, 
pretty much forces us to exclude hash-based join algorithms. 

Do we really care about ordering in this case? For "human-readable" documents,
certainly. For "data-oriented" documents we don't; specifying
ordering forces us to use a, sometimes slower, nested-loops join implementation.

> As far as XQuery being more
> "succinct" than XSLT, that's due to XSLT's XML syntax, not due to its
> declarativity.

I agree. But, when needed, XQuery will also be representable in XML! 
The "XML Query Requirements" document mentions (section 3.2):

"One query language syntax MUST be expressed in XML in a way that 
reflects the underlying structure of the query."

The advantage of XQuery over XSLT is that humans will still be able to write 
and reason about queries in the cleaner, more succinct format.

Best regards,

Vassilis Papadimos.