Newsgroups: alt.usage.english,sci.lang
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!zombie.ncsc.mil!news.duke.edu!news-feed-1.peachnet.edu!gatech!howland.reston.ans.net!ix.netcom.com!netcom.com!smryan
From: smryan@netcom.com (Principal Pooper)
Subject: Re: Chomsky's Theorem
Message-ID: <smryanD5tn7w.5D6@netcom.com>
Followup-To: alt.usage.english,sci.lang
Organization: Morse Science High
X-Newsreader: TIN [version 1.2 PL1]
References: <3kjtta$a57@riscsm.scripps.edu>
Date: Wed, 22 Mar 1995 03:04:44 GMT
Lines: 27
Sender: smryan@netcom2.netcom.com

While CNF is necessary for a CKY parser, a GHR or Earley
parser does not need such preprocessing, and so CNF is
not as important as it was.

What's more important for computer languages is that
LL(k) < LL(k+1) < LR(k) = LR(1).

--------------------------------------------------------

Discarding formal linguistics is a major step, not to be
done lightly. Since a type 0 language is equivalent to a
Turing Machine and, by the CT hypothesis, equivalent to
any formal symbol manipulation, the claim that languages
cannot be described by a formal grammar is a strong claim
that humans transcend Turing Machines. And thus AI is
impossible. 

Not that you can't claim it, but it should not be done
trivially or carelessly.

And since the strong CT hypothesis is neither proved nor
disproved, don't demand people accept your claim.
-- 
The pair depart upon a horse            | smryan@netcom.com     PO Box 1563
and fare to face their future's course. |             Cupertino, California
Away! the walls of weighted stone!      | (xxx)xxx-xxxx               95015
Away! the wealth and worried throne!    |    What happened to little Artie?
