>
>Exactly.
>
>Nothing new here. I was partner in Savvy Inc., which owned rights to a 12 state territory to market and support SAVVY, an "AI" using natural langauge.
>Requests like "Give me a list of all customers in georgia who have purchased less than $2,000 of widgets in the last six months, sorted by zip." Pretty amazing in 1982 -83.
>I was a perhipheral card on the Apple and a Floppy disk on the IBM PC. If it didn't understand a word, or you mispelled, it would offer you choices on what it "thought" you meant. When you responded it created more AI links on the response to that choice.
>
>It became more obvious that untrained folks did not understand how to phrase a question to get the computer to supply the info they wanted. Even that simple task takes training. That simple example above doesn't begin to show the complexities in a multi-bounded query. Even VFP6's query parser barfts if things get even slightly complicated.
>JLK


I think there's lots of potential for English Query. Too bad it's soooo hard to set up.