I've used pastebin to share my dataset.

But, also, when isolating problems in my AoC 11 implementation, I
found it useful to use a routine to help isolate the first place where
a discrepancy showed up.

(And then again to trim off a front section from the sequence until
just before the discrepancy vanished.)

That turned out to be really useful for getting a discrepancy in a
data set which I can personally inspect.

[My original implementation used geometric distance instead of purely
counting steps, and I didn't realize that that was a mistake until I
could see how it was giving bad answers for the second part - with
this minimal example I could see that I was getting a second part
answer which was smaller than my first part answer.]

Thanks,

-- 
Raul


On Sat, Dec 16, 2017 at 3:32 PM, Brian Schott <[email protected]> wrote:
> Did you notice my comment about an anomaly using i. to find the first
> <'N' ? If that is really a bug, then maybe it is producing some of the
> peculiarity.
>
> When AoC offers the ability to share a question, does it mean that the
> identical data set would be shared, too? If so, that might be an easy
> way to share my data. Otherwise, I could post it somewhere like dpaste.com .
>
> Thanks for your help,
>
> On Sat, Dec 16, 2017 at 3:24 PM, Raul Miller <[email protected]> wrote:
>
>> So, looking further at this, I ran your expression (for test=: ...)
>> against my aoc dataset (then translated it to lower case and parsed it
>> using ;: so that it's in the format that my routines expected).
>>
>> The original data gave me a path with 423 ne and 241 se.
>>
>> The result of your test expression gave me a path with 144 n and 664
>> ne which meant, I thought, that you have at least one bad substitution
>> somewhere.
>>
>> But then I ran the same procedure again against the same data set and
>> got a completely different result (which also matched with my answer).
>>
>> So I don't know what's going on.
>>
>> Good luck,
>>
>> --
>> Raul
>>
>>
>>
>>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to