NPR puzzle

[Jim Hurley]

>
>Message: 12
>Date: Fri, 22 Jul 2005 08:04:53 -0700
>From: Geoff Canyon <gcanyon at inspiredlogic.com>
>Subject: Re: NPR puzzle
>To: How to use Revolution <use-revolution at lists.runrev.com>
>Message-ID: <CE0BDA59-691C-4540-B44B-D25AC8286C73 at inspiredlogic.com>
>Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
>
>The engine optimizes the tests in conditionals. So in this statement:
>
>if a=3 and b=2 then...
>
>b never gets evaluated if a is not 3. I rewrote Dick's very nice code 
>using this principle, to get rid of the nested if statements. 
>Interestingly, my code slowed down a couple milliseconds. Just from 
>13 to 15, but I wonder why that might be. Maybe the engine goes to 
>the trouble of getting the substrings before doing the tests.
>
>Here's the code:
>
>    put the milliseconds into tMilliseconds
>    repeat for each line tSymbol in field 2
>      put "true" into S[tSymbol]
>    end repeat
>    repeat for each word W in field 1
>      if S[char 1 to 2 of W] and \
>          S[char 3 to 4 of W] and \
>          S[char 5 to 6 of W] and \
>          S[char 7 to 8 of W] and \
>          S[char 9 to 10 of W] then \
>          put W & return after tHits
>    end repeat
>    put the milliseconds - tMilliseconds into tElapsedMilliseconds
>    put number of lines in tHits && "hits in" \
>        && tElapsedMilliseconds && "milliseconds" \
>        & return & return & tHits


Geoff,

It appears that RR also does not continue with the nested ifs after 
it gets a false. I put a "beep" after the next to last if in Dick's 
handler and the handler only beeped once.

Jim