RE: drop sortby from simplesearch

> 4) In order to describe sort in query schema discovery, we 
> were forced to
> introduce an artificial property (dav:rank) (see 7.19.1)  
> this is ugly, and
> we can get rid of it by abolishing sort.

I would quibble on the reason: We didn't have to introduce
dav:rank to "describe sort". Dav:sort is not mentioned
in sections 7.7*. Also, having a rank does not imply
the server must sort on it. (In fact there are important
systems where the client can do the sort.) Also, we can
get rid of dav:rank without getting rid of sortby, if
we choose to.

But that's just quibbling. I have a more important problem with 
dav:rank as currently specified, and that is that the spec.
says: "The data type is not defined." The datatype would
have to be stated in the result of the QSD. Therefore,
I propose striking that sentence. 

I further propose making the ordering of 
relevance explicit: For example, is rank 1 
the most or least relevant hit? I propose we say 
that as dav:rank increases, the relevance increases.
The client can not do a relevance sort without knowing the
datatype and whether larger values are more relevant
than smaller values.

If we don't do these two things, then we should give up on our
attempt to standardize the rank property and drop it.

Whatever we do or don't do will not prohibit collections from 
advertising whatever rank property they choose. It would 
be nice if the QSD had an optional narrative description 
element for each property, so user's wouldn't have to guess 
what the property is good for from its name. That could
help with all properties, not just rank properties.

Alan Babich

Received on Friday, 24 July 1998 15:42:09 UTC