Newsgroups: rec.arts.int-fiction
Path: nntp.gmd.de!news.ruhr-uni-bochum.de!news.rwth-aachen.de!uni-paderborn.de!fu-berlin.de!main.Germany.EU.net!EU.net!nntp04.primenet.com!nntp.primenet.com!howland.erols.net!netcom.com!erkyrath
From: erkyrath@netcom.com (Andrew Plotkin)
Subject: Re: [Inform] Parsing Ambiguity???
Message-ID: <erkyrathDws44x.I6t@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
X-Newsreader: TIN [version 1.2 PL1]
References: <4vfkm2$qne@news-e2d.gnn.com> <erkyrathDwI6vp.Cyt@netcom.com> <322105D6.46CB@cosc.canterbury.ac.nz>
Date: Tue, 27 Aug 1996 04:22:09 GMT
Lines: 33
Sender: erkyrath@netcom18.netcom.com

Greg Ewing (greg@cosc.canterbury.ac.nz) wrote:
> It would be nice if the parser understood the notion of
> referring to an object which is a part of another object,
> so that you could just call the parts "window" or "door" and
> be able to say

>   open window of car

> Even better if it would also accept

>   open car window

> as meaning the same thing.

> This would entail making two attempts at finding an object
> to match each noun phrase, the first time looking for a
> single object matching all the words, and if that fails,
> looking for a hierarchical match.

Well, if you name the window "window of car" you get that effect (since 
the Inform parser ignores the order of nouns in an object's name.)

To set it up automatically (for certain objects) would be pretty easy; 
adjust the Refers procedure to scan the parents of an object as well as 
the object itself for names. You would not want this to happen in most 
cases, though. Hm. Maybe.

--Z

-- 

"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the
borogoves..."
