[seqfan] Re: Measuring OEIS Sequences ?

Charles Greathouse charles.greathouse at case.edu
Mon Dec 19 21:26:10 CET 2016


Yes, I think it's reasonable to link to data repositories. Certainly many
existing sequences reference CODATA.

In the particular case of A038534 and A038533 you should probably include a
descriptive comment since the link may not otherwise be clear to readers.

If you'd like to write an article on the OEIS wiki, feel free.

I don't think there are any requirements on the precision for external data
sources. When the external source *is* the sequence, at the very least we
should not include digits which are entirely unknown. If a constant is
given as 1.234567(11) meaning a 1 standard deviation range from 1.234556 to
1.234578, only the digits 1, 2, 3, 4, 5 should be given, certainly not the
6 or 7.

Charles Greathouse
Case Western Reserve University

On Fri, Dec 16, 2016 at 6:41 PM, Brad Klee <bradklee at gmail.com> wrote:

> Hi Seqfans,
>
> Some of the OEIS sequences are measureable in nature, a premiere example is
> the pair ( A038534 , A038533 ), which together describe the energy
> dependence of a plane pendulum's period.
>
> For these types of sequences, we could link to measurement repositories,
> where available. See for example:
>
> https://github.com/bradklee/DataAnalysis/tree/master/PlanePendulum
> >>
> https://github.com/bradklee/DataAnalysis/blob/master/
> PlanePendulum/ExamplePlot.png
> >>https://github.com/bradklee/DataAnalysis/blob/master/
> PlanePendulum/fit.log
>
> Already in v0.1, the fit log gives the following estimates for extracted
> parameters
>
> (a(1),a(2),a(3) ) = ( 0.2448 +/- 0.0045 , 0.1635 +/- 0.0113 , 0.0934 +/-
> 0.0152 )
>
> to  compare with theoretical frac. values ( decimals truncated )
>
> (a(1),a(2),a(3) ) = ( 1/4, 9/64, 25/256 ) ~ (0.2500 , 0.1406 , 0.0976)
>
> All measurements are within 3 sigma, and the percent difference is
> approximately 2%, 16%, 4% for linear, quadratic, cubic coefficients
> respectively.
>
> So a few questions:
> ~ Should physically measureable sequences have there own tag / wiki topic ?
> ~ Should we link to data repositories if relevant / available ?
> ~ If yes, should we define accuracy / precision requirements ?
>
> For the record: I think Y / Y / Y . Let's see if anyone has an argument for
> N / N / * .
>
> Best Regards,
>
> Bradley Klee
>
> In the fanfiction voice:
> <<Throughout time the Force remains cloaked in antinomies
> where opposites rarely find themselves exactly equals
> and the power strugle pertains to nothing less valuable
> than fundamental laws of universe, cast in mortal form,
> in thrift teaching. With satisfactory accuracy and precision
> data slices defend the most adherent and blasted derision.>>
>
> --
> Seqfan Mailing list - http://list.seqfan.eu/
>


More information about the SeqFan mailing list