So, it's more a problem of digits being lost somewhere rather and not using a
suitable compaison tolerance? Did you have to sequentially test all
possibilities from the default upwards, or did you analytically come to 1e-11
somehow?
May 17, 2022, 17:44 by hauke.r...@uni-jena.de:
> but with
> (9!:19) 1e_11
> I get
> ({.e.}.)"1 w
> 0 0 0 1 1
>
>
> Am 17.05.22 um 16:40 schrieb Hauke Rehr:
>
>> oops, ignore that one – I see my mistake now
>>
>> Am 17.05.22 um 16:37 schrieb Hauke Rehr:
>>
>>> I didn’t revisit pe80 (its been quite some time I last looked at pe) but
>>>
>>> I get
>>> <./ ({.>./@:|@:-}.)"1 w
>>> 0.539284
>>> so you would need a tolerance greater than that
>>> if you want to get at least one 1
>>>
>>> with a sane value for tolerance, a 1 is not to be expected
>>> so what do you want to achieve, what do you expect and why?
>>>
>>>
>>> Am 17.05.22 um 16:22 schrieb 'Viktor Grigorov' via Programming:
>>>
>>>> Hey,
>>>>
>>>> I'm working on project euler 80, and I'm getting stack errors. Tracing the
>>>> values, step by step, everything is in order, other than the comparison. I
>>>> have the following:
>>>>
>>>> ]w=:{{ ((1&|@%@{.),])y }}^:(>:i.5) @(1&|)@%: 23
>>>> 0.256547 0.795832 0 0 0 0
>>>> 0.897916 0.256547 0.795832 0 0 0
>>>> 0.11369 0.897916 0.256547 0.795832 0 0
>>>> 0.795832 0.11369 0.897916 0.256547 0.795832 0
>>>> 0.256547 0.795832 0.11369 0.897916 0.256547 0.795832
>>>>
>>>> ({.e.}.)"1 w
>>>> 0 0 0 0 0
>>>> 9!:19 0.00001
>>>> 9!:19 1.00000000000000008e_5
>>>> ({.e.}.)"1 w
>>>> 0 0 0 0 0
>>>>
>>>> Lowering the comparison tolerance further doesn't help. What am I doing
>>>> wrong? This can't be a 904 issue, since it happens on the 9.03.08 as well.
>>>> ----------------------------------------------------------------------
>>>> For information about J forums see http://www.jsoftware.com/forums.htm
>>>>
>
> --
> ----------------------
> mail written using NEO
> neo-layout.org
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
