Kristjan Onu wrote:
Raymond E. Rogers wrote:
Either you or I could add a ellint_[KE]comp comment to the ellipke
code. I could enter the k^2/m directions in the ellint_[KE]comp
documents; or perhaps that would go in both cases?
I think there is often confusion regarding whether elliptic integrals
are defined with their input argument squared or not, ie. whether it's
E(k) = \int_0^\pi/2 dt \sqrt((1 - k^2 \sin^2(t)))
or
E(k) = \int_0^\pi/2 dt \sqrt((1 - k \sin^2(t)))
With a little effort, people using either ellint_[KE]comp or ellipke
will find how each function is defined, but I think we should decide
whether it's worth explicitly documenting the definitions used within
each function. To me the arguments in favor of adding this
documentation are: (i) it's a common point of confusion, (ii) that
among Octave-Forge functions different definitions are used. An
argument against adding this documentation might be that giving
references to Abramowitz and Stegun and the GSL documentation is
sufficient.
Okay I will edit in some changes and submit your program. The changes
would be adding to the corrections section, adding the equation in tex
form in the comments, adding some text about m=k^2, breaking out the
test section to test_ellipke.m . If you think it's worthwhile I could
add the GSL tests to the test file while I am at it. I did this for the
hyperg_[M,U].
First I will get the current svn:)
RayR
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Octave-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/octave-dev
