sam clearly explains why named parameters and namespaces are good when defining interfaces. one thing i'd add is that i absolutely loathe the use of things that look like a perfectly valid url as a more general uri. when something starts with 'http://', i expect to be able to do an HTTP GET on that resource. identifiers should use a scheme other than 'http' when that is not the intended use.

(and yes, i realize that i appear to be in the hopeless minority (or very silent majority) on this, and probably attach more significance to the difference between a uri and a url than i should. but it's just the sort of thing that creates enough cognitive dissonance for me to run away from w3c specifications.)