> Date: Wed, 15 Dec 2004 23:15:31 +0000
> From: Dr Robert Sanderson <[log in to unmask]>
>
> >> I think that's exactly what cql.anywhere means.  Search all
> >> indexes that you know about, but you can't be expected to search
> >> indexes that you don't
> >
> > Well, I will leave it to the profile author to say whether that
> > was what he meant.  I read the Adlib-specific definition to mean
> > that they wanted "anywhere" to mean "in any of the Adlib access
> > points only".
>
> Then I would suggest:
>
> adlib.anywhere   Search in any of the indexes defined in the adlib context
>                   set.

Sounds right to me -- if that's what the writer did mean.

> > For Adlib, there _is_ an implementation difference between (the
> > putative) cql.record access point on cql.anywhere. in that the latter
> > makes use of all the indexes that are to hand, but the former makes a
> > physical scan of all records, so it can even find the term in fields
> > that are not indexed.  As I implied, not every server will implement
> > this (which is fine) but it does seem like a useful distinction for
> > those that do.
>
> cql.anyField ?  I don't like cql.record as a name, as to me that would
> imply that a string search should be matched against the entire record,
> rather than any subfield within the record.

Agreed -- your name is better,

> Also, if the text isn't indexed at all, how do you determine where to
> start matching and where to end?

I don't understand.  Whatever you might otherwise have done at
indexing time, you just do at search time instead.  Run-time
performance will suffer, but the pay-off is of course smaller index
files.

> >>> This is _not_ what "=" means in CQL.  It means that the term is
> >>> word-structured, irrespective of the index being searched, unless
> >>> overridden by a relation modifier.
> >>
> >> I disagree, Mike.  It means, currently, exact equality, but does not
> >> say how equality is to be determined.
> >
> > Au contraire.  This is precisely the difference between "=" and
> > "exact".
>
> =/cql.number  must surely mean exact numeric equality?

Er.

Oh, look!  A badger!

 _/|_    _______________________________________________________________
/o ) \/  Mike Taylor  <[log in to unmask]>  http://www.miketaylor.org.uk
)_v__/\  "When you are young, you enjoy a sustained illusion that
         sooner or later something marvelous is going to happen [...]
         But it isn't like that at all.  But if it isn't, where did
         the idea come from?" -- Brian Aldiss, _Helliconia Summer_

--
Listen to free demos of soundtrack music for film, TV and radio
        http://www.pipedreaming.org.uk/soundtrack/