Re: Query fallback critical

>Maybe it would be better not to count on queries yet and 
>accept that the Web can't do better than send the whole
>instance, for client-side processing of queries.

Right. Indeterministic results are useless for addressing
purposes, which is the point I've been trying to make.

The only way I know of getting deterministic addressing is
if the server controls the address space (ie. it generates
it), *or* if there is a standard addressing syntax (which
I would like to see, but think adoption unlikely, as I
proposed it before, and it failed to catch on).

This is not to say that queries are any more or less useful 
than current WWW queries. By definition, they are not.