> If we keep word lists, then we need some mechanism to indicate what kind of
> word list they contain.  I much prefer the more extensible proposal of
> following the index name with a colon separated list of modifiers.  I am
> dead set against a bunch of special characters with limited extensibility
> and even more limited understandability.

How is:
        indexset.index:token = "term"
any less extensable than:
        indexset.index token "term"

You still have to define a set of tokens.  I would much prefer a space
separation rule between index, relation and term in CQL (obviously in XCQL
it's unnecessary) and then allowing people to extend the tokens used for
relation.

If you want to have a relevance search, what is better about:
        dc.description:relevance = "dublin core"
compared to:
        dc.description relevance "dublin core"

(or our shorthand for it of =~)

If > is still a relation like =, then what is the meaning of:
        dc.description:relevance > "dublin core"
?
We would need a whole host more diagnostics along the line of 'Invalid
relation and index modifier combination' ... Urgh.


The one useful aspect that I could think of is if you wanted to qualify a
different operator than '=' ... for example >=, but can anyone give a
concrete example of when this would actually be useful?

Or would = and > become modifiers? If they're not modifiers, then what is
the difference between them such that some are modifiers and some are
relations?

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