in reply to Re^3: Challenge: prefix($x, $y) . suffix($x, $z) eq $x
in thread Challenge: prefix($x, $y) . suffix($x, $z) eq $x

shmem,
This doesn't seem to scale very well. I started to run it against a 60K word dictionary file and I gave up waiting for it to produce output after a few minutes. Perhaps you could modify it to accept a target word from the command line and only do all matches if one isn't provided?

Cheers - L~R

  • Comment on Re^4: Challenge: prefix($x, $y) . suffix($x, $z) eq $x

Replies are listed 'Best First'.
Re^5: Challenge: prefix($x, $y) . suffix($x, $z) eq $x
by shmem (Chancellor) on Apr 22, 2009 at 23:59 UTC

    Limbic~Region,

    well, it was never meant to scale ;-)
    I've updated my post. Running it with 'ivanova' as argument now yields 

    real    0m0.067s
user    0m0.051s
sys     0m0.003s

    [download]

    on average.

    --shmem

[reply]
[d/l]
      shmem,
      Much better - but it helped me find another bug in your code. 
      $perl shmem.pl hello
found: prefix(haughtiness humanoid ... hello

      [download]

      prefix(hello, hello) ne 'h' for a number of reasons. First, they have more in common then h. Second, it would consume the entire word - both of which are against the rules.

      Cheers - L~R

[reply]
[d/l]

        No, Sir.

        It is not prefix(hello, hello), but prefix(haughtiness, hello) etc. haughtiness, humanoid and hello have the prefix 'h' in common, and the 'ello' is provided by the words output in the suffix section. Run the prefix() sub passing it (qw(hello hello)) and you will get no return value.

        So (humanoid hello, haughtiness, cello) is a valid tuple. The presence of 'hello' in the prefix section is an artifact of the implementation... er, bug ;-)

        --shmem

[reply]
[d/l]
[select]

In Section Meditations