if the original numbers have "full" precision (instead of 4 decimals) J gives
the result you want
x: % p: 1 + i.10
1r3 1r5 1r7 1r11 1r13 1r17 1r19 1r23 1r29 1r31
x: % +: p: 1 + i.10
1r6 1r10 1r14 1r22 1r26 1r34 1r38 1r46 1r58 1r62
On Thursday, August 1, 2019, 05:01:44 p.m. EDT, Nimp O <[email protected]>
wrote:
Ah, so it appears there is a very low global upper bound (2^_34) on the
tolerance of all verbs that can use Fit.
Perhaps this upper bound could be removed, at least for x:, since I believe the
main use cases for tolerance on this verb would be bigger values, as
illustrated in my earlier example.
At least in this context, I can't see anything catastrophic happening by
increasing this tolerance, on the contrary.
But maybe this change cannot be made in such an isolated way.
Thanks.
________________________________
From: Programming <[email protected]> on behalf of Raul
Miller <[email protected]>
Sent: Thursday, August 1, 2019 11:12 AM
To: Programming forum <[email protected]>
Subject: Re: [Jprogramming] Extended precision tolerance
Looks like this specifies comparison tolerance.
https://code.jsoftware.com/wiki/Essays/Tolerant_Comparison
"Appendix C, System Conventions and Limits, says that the tolerance is
less than or equal to 2^_34 . Historically, the upper bound on the
tolerance was chosen so that comparisons involving 32-bit integers are
exact (as if the tolerance is 0)."
FYI,
--
Raul
On Thu, Aug 1, 2019 at 9:33 AM Nimp O <[email protected]> wrote:
>
> Hello everyone,
>
> What is the use of non-zero tolerance for the x: verb?
>
> A little example:
>
> a=:0.1667 0.8333 0.4286 0.1111
>
> x:a
> 1667r10000 8333r10000 2143r5000 1111r10000
>
>
> In Mathematica:
> In[1]:= Rationalize[{0.1667, 0.8333, 0.4286, 0.1111}]
> Out[1]= {1667/10000, 8333/10000, 2143/5000, 1111/10000}
>
> In[2]:= Rationalize[{0.1667, 0.8333, 0.4286, 0.1111}, 0.01]
> Out[2]= {1/6, 5/6, 3/7, 1/9}
>
> Back to J:
> I was expecting the that Fit could be used the same way, as in a tolerance
> for the rationalization, but I get:
>
> x:(!.0.01) a
> |domain error
> | x: (!.0.01)a
>
> I was waiting for this expression to return my desired output:
> 1r6 5r6 3r7 1r9
>
> Thanks.
>
>
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
