Newsgroups: rec.arts.int-fiction
Path: gmd.de!jvnc.net!yale.edu!qt.cs.utexas.edu!cs.utexas.edu!zaphod.mps.ohio-state.edu!malgudi.oar.net!caen!uunet!hobbes!jondr
From: jondr@sco.COM (Karen Silkwood's car)
Subject: Re: What words to use and recognize
Organization: The Santa Cruz Operation, Inc.
Date: Fri, 18 Dec 1992 17:39:49 GMT
Message-ID: <1992Dec18.173949.837@sco.com>
References: <BzDJLv.8xn.2@cs.cmu.edu>
Sender: news@sco.com (News admin)
Lines: 25


pww+@A.GP.CS.CMU.EDU (Peter Weyhrauch) writes:
>A case in point is the Oz system.  We currently generate all our
>narrative text, and soon will generate character dialogue as well.
>Since the same lexicon is used for generation and parsing, any word
>used in generation is parseable.  The grammars are currently
>different, but encode (roughly) the same functionality.  Notice that
>you could use a more universal parser if you chose.  The perhaps
>unacceptable downside is that our output is uninspired for now.  Part
>of our research efforts are aimed at doing stylistic generation.  For
>your curiousity, I'll attach a fragment at the end of my post.

interesting research and no doubt leading towards a satisfying conclusion
but i guess i'm a bit of a conservative when it comes to IF.  i had many
many hours of entertainment playing infocom games as a teenager and i have
no problem with games coming out nowadays that live up to the standard set
by infocom.  to wit: imaginative, interesting prose and a sometimes
frustrating parser.  if that means i'll be pulling my hair out for a few
hours trying to figure out why PUSH SATCHEL IN FRONT OF PANEL doesn't work
while COVER PANEL WITH SATCHEL does, then so be it.

-- 
Jon Drukman (God's personal DJ)                 uunet!sco!jondr   jondr@sco.com
-------------------------------------------------------------------------------
I was an infinitely hot and dense dot.
