Duplicated titles (e.g 305 and 306) would make
a difference in the result if you encoded titles
as integers. Then 305 306 would look like a
legitimate chain with no repetitions.
The curveballs could be deliberate. The ITA thing
is not a contest but a job application screener
("solve this puzzle; get a job offer"). Perhaps
part of the test is to see how you deal with
real world inputs.
----- Original Message -----
From: Joey K Tuttle <[EMAIL PROTECTED]>
Date: Tuesday, May 22, 2007 3:40 pm
Subject: Re: [Jprogramming] Sling Blade Runner
> At 13:28 -0700 2007/05/22, Roger Hui wrote:
> >My current champion has length 278 (titles).
> >
> >A few key insights made possible several orders
> >of magnitude improvements in the processing speed.
> >
> >Also, beware that the input file has some subtle
> >flaws which may make a difference (some lines
> >terminate in CR instead of CRLF; duplicate lines).
> >
>
> Actually, not too subtle. I noticed the first time
> I looked at the file that it seemed to be CRLF
> delimited so, as my usual practice did CR-.~ fread
>
> I didn't notice that there were 1774 fewer CRs than
> LFs ... and that there were 8 duplicated titles -
> although that shouldn't make any difference in the
> results. Still, one would hope that posted contests
> would take some effort to preen sample files...
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
