From rob@twcny.rr.com Fri Mar 15 15:15:04 2002
Return-Path: <rob@twcny.rr.com>
X-Sender: rob@twcny.rr.com
X-Apparently-To: lojban@yahoogroups.com
Received: (EGP: unknown); 15 Mar 2002 23:15:04 -0000
Received: (qmail 38002 invoked from network); 15 Mar 2002 23:15:03 -0000
Received: from unknown (216.115.97.172)
  by m2.grp.snv.yahoo.com with QMQP; 15 Mar 2002 23:15:03 -0000
Received: from unknown (HELO mailout6.nyroc.rr.com) (24.92.226.177)
  by mta2.grp.snv.yahoo.com with SMTP; 15 Mar 2002 23:15:03 -0000
Received: from mail1.twcny.rr.com (mail1-1 [24.92.226.139])
  by mailout6.nyroc.rr.com (8.11.6/Road Runner 1.12) with ESMTP id g2FNF2u08476
  for <lojban@yahoogroups.com>; Fri, 15 Mar 2002 18:15:02 -0500 (EST)
Received: from riff ([24.92.246.4]) by mail1.twcny.rr.com
  (Post.Office MTA v3.5.3 release 223
  ID# 0-59787U250000L250000S0V35) with ESMTP id com
  for <lojban@yahoogroups.com>; Fri, 15 Mar 2002 18:15:01 -0500
Received: from rob by riff with local (Exim 3.33 #1 (Debian))
  id 16m0ub-0000UJ-00
  for <lojban@yahoogroups.com>; Fri, 15 Mar 2002 18:15:01 -0500
Date: Fri, 15 Mar 2002 18:15:01 -0500
To: lojban@yahoogroups.com
Subject: Re: [lojban] ChomskyBot
Message-ID: <20020315231501.GB1859@twcny.rr.com>
References: <3C927952.20005@cs.technion.ac.il> <Pine.GSO.4.40.0203151552540.28082-100000@ucsub.colorado.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.GSO.4.40.0203151552540.28082-100000@ucsub.colorado.edu>
User-Agent: Mutt/1.3.24i
X-Is-It-Not-Nifty: www.sluggy.com
From: Rob Speer <rob@twcny.rr.com>
Reply-To: rob@twcny.rr.com
X-Yahoo-Group-Post: member; u=2572649
X-Yahoo-Profile: squeekybobo

On Fri, Mar 15, 2002 at 03:56:50PM -0700, Jay Kominek wrote:
> Richard Curnow also has a system under development.
> 
> All of which are quite a bit more complicated than the ChomskyBot.
> 
> (Mine use an algorith from a fellow out of .nz which analyzes the grammar
> to determine all the proper weightings so you get sentences of exactly the
> length you want. Nora's (the LLG one) just has hard coded weights. Not
> sure what Richard's does, though I think he told me at one point.)

How hard would it be to modify yours to not use that algorithm? I'd like
a random sentence generator that works on Linux, and while yours has a
good idea behind it, the output leaves something to be desired. The vast
majority of seven-word sentences, for example, consist of seven
unrelated cmavo, indicating that weighting really is important.

Heck, I'd fiddle around with the weights if I had some way to adjust
them.

-- 
la rab.spir
noi sarji zo gumri


