On Sat, 7 Sep 2002, Ray denenberg wrote:
> From: "Robert Sanderson" <[log in to unmask]>
> > > Just as importantly, when we move to support Scan, we aren't going to
> have
> > > any truncation characters to help.  So when I scan on Title="word", do I
> > Yes.  This is, IMO, the killer argument against combining Title and
> > TitleWord

> If it's the latter: we haven't begun to define the scan service. We can
> define it  however we want and include whatever parameters that are
> appropriate.

I understood the latter.

Assume that there's one 'title' index.  You want to scan for keywords.

So you do something like:
<scanRequest>
  <index>title</index>
  <term>a</term>
  <numRequested>20</numRequested>
  <stepSize>0</stepSize>
  <startPosition>1</startPosition>
  <structure>keyword</structure>
</scanRequest>

And it gives you back 'aardvark' 'absynth' 'accelerate' ...

Then the query, I would assume to be: title="|aardvark|" ?

This would then necessitate a list of supported structure attributes per
index, one of which would be the default.  This would be easy enough to do
in Explain, but seems very classic Z39.50 (not that that's a bad thing of
course).

Bah. I'm reconvinced. I'd rather add a one line '<supports
type="structure">keyword</supports>' to each index than duplicate it and
tack 'Word' on the end of the name. Then it's up to the server to use the
most appropriate method to fulfil the query rather than the client telling
the server which index it thinks it wants based on Explain.

I'm sure there's a nice American expression for going backwards and
forwards from one side to the other, but I can't think of one at the
moment :)

Rob

--
      ,'/:.          Rob Sanderson ([log in to unmask])
    ,'-/::::.        http://www.o-r-g.org/~azaroth/
  ,'--/::(@)::.      Special Collections and Archives, extension 3142
,'---/::::::::::.    Twin Cathedrals:  telnet: liverpool.o-r-g.org 7777
____/:::::::::::::.              WWW:  http://liverpool.o-r-g.org:8000/
I L L U M I N A T I