Re: [FWP] A "Cheating" hangman dealer

[Ronald J Kimball <rjk@linguist.thayer.dartmouth.edu>]

On Mon, Feb 19, 2001 at 03:53:18PM +0100, Jerome Quelin wrote:
> The output of diff on your fixed program:
> 60c60
> <         $word =~ s/[^$guess]/./g;
> ---
> >         $word =~ s/[^$guess]/[^$guess]/g;     
> 

Thank you very much for the patch!  Very foolish of me not to notice that
problem.


> Rich Morin wrote:
> > Fun to play, too, though I found that I had no chance of winning
> > against it unless _I_ cheated, as well!
> 
> Well, it's not true. The program is quite stupid. If you
> guess letters in always the same order, and pick up letters
> in a careful order (so that you eliminate as much words as
> possible at each pick), you will end up always to the same
> word!
> If you use the /usr/dict/words file, try this sequence:
> a,e,i,o,u. You will end up with only two words: 'rhythm' and
> 'nymphs'. Simply pick the letter 'y', and you will know
> which word was picked. You can now win easily, since you
> know the word to find.

Heh.  I tried that with the ENABLE word list, and found that it contains
the word 'tsktsk'.  :)

BTW, note that the contents of the /usr/dict/words file is system
dependent.  (Actually, so is the file's location.  Any program which uses a
word list should have an option for specifying an alternate path to the
word list file.)


> Rich Morin wrote:
> > I also found that I had to cheat (i..e., peruse Camel III) to
> > figure out what was going on, particularly in partition_words. 
> 
> You can easily skip the partition_words sub. Perl provides
> the grep function that is more convenient.

But less memory efficient, because it builds a copy of the original array.
partition_words modifies the array in place.  I wouldn't be surprised if
the grep were faster, though.


Ronald

==== Want to unsubscribe from Fun With Perl?  Well, if you insist...
==== Send email to <fwp-request@technofile.org> with message _body_
====   unsubscribe