Path: news.net.uni-c.dk!newsfeeds.net.uni-c.dk!newsfeed1.uni2.dk!news.get2net.dk!not-for-mail From: "Jim Lucas" Newsgroups: comp.ai.neural-nets,comp.lang.apl,comp.lang.awk,comp.lang.beta,comp.lang.cobol,comp.lang.dylan,comp.lang.forth References: <3AABD82E.C4D4E5CB@Home.Com> <4rVq6.559$ou4.12675@news.get2net.dk> <3AAC3139.5703EFD1@Home.Com> Subject: Re: Einstein's Riddle Lines: 72 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.2919.6600 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600 Message-ID: <711r6.40$fo5.7330@news.get2net.dk> Date: Mon, 12 Mar 2001 10:47:11 +0100 NNTP-Posting-Host: 194.234.11.121 X-Complaints-To: abuse@danbbs.dk X-Trace: news.get2net.dk 984390851 194.234.11.121 (Mon, 12 Mar 2001 10:54:11 MET) NNTP-Posting-Date: Mon, 12 Mar 2001 10:54:11 MET Organization: DanBBS Xref: news.net.uni-c.dk comp.ai.neural-nets:67530 comp.lang.apl:29367 comp.lang.awk:17091 comp.lang.beta:12738 comp.lang.cobol:102491 comp.lang.dylan:24150 comp.lang.forth:78501 "David Ness" wrote in message news:3AAC3139.5703EFD1@Home.Com... > Jim Lucas wrote: > > > > You appear to be making an interesting claim, i.e., that the range of > > techniques for solving this puzzle computationally is limited (and also too > > simple to be interesting?), regardless of the language. Certainly, simple > ... > > No. I wasn't making any claim, other than... I disagree. By praising the New Scientist puzzle for "exposing lots of different and contrasted technique" and characterizing the "Einstein" puzzle as "mediochre" [sic] and unable to "accomplish" the same variety of technique, you have indeed made a claim, even if it was by implication. > ...to say that I've done enough of the kind > of puzzle suggested to lead me to believe > that answers were likely to be boring,... Answers to the puzzle as stated, or answers in the sense of programming solutions? I'm not expert at such logic problems, but I found it easy enough with pencil and paper to determine that the German owns the fish (assuming that the problem is intended to be one of logic, and not one of linguistic quibbles). I'm not convinced that it will be as easy to write a computer program to do it for me, though I would hope that such a program would also be able to handle more "difficult" versions of the same type of problem. But in all your experience with solving this type of problem, you haven't said whether you've ever written a program to do it. It's the programs that I'm curious about. > ...and would mostly involve discussion about > modelling the (human) language that was used > to present the problem. Yes, that *is* the interesting part: how to specify and process such conditions as "next to"; how to apply the different sorts of conditions; how to select which conditions to apply first in order to maximize efficiency. Maybe it's simple, but maybe not to someone who hasn't tried it before. And is there really only one way to do it? At least in APL, there are certainly multiple ways to eliminate duplicate rows from a matrix, which I think is logically a much simpler "puzzle". > So far almost all of the responses that have > appeared have suggested that this particular > point is true. There is argument about `owning > fish'. We have the possible answer `none'. > There can be argument about whether `first > house' means left-most or right-most. Stuff > like that. So far, most of the responses have *not* been about *modelling* the language, but about quibbling with the semantic interpretation, where the *intended* semantics are quite clear. They have been attempts to avoid the real, intended puzzle. I can only wonder if that's because it's *not* so easy to write a program to solve that sort of puzzle. I find it interesting that John R. Clark indicates that *his* solution "would be too long to post here." > Since I believe my own point, I'm not > interested in playing with the problem, > so I demur from worrying about a solution. Fair enough, but *I* hope that others *will* submit solutions. I even hope that I'll find some time to try it, since I haven't previously tried programming this kind of problem. /Jim