Proposal for naming of Bessel functions



Commercial Macsyma seems to use Bessel_J[n](a)  or maybe bessel_j[n](a)
depending on the setting of  display_case:initial_caps.

There is a convention that people separate long identifiers by
UsingCapitals or using_underscores.  This does both.  Also note
that Macsyma uses [n] for the order.  Subscripted functions.
Not like some other systems.
RJF




Raymond Toy wrote:

> While adding the numeric routines for the bessel functions, it
> occurred to me that what maxima has is somewhat messy.  There's j0(x),
> j1(x), and jn(x, n) which compute the Bessel functions for order 0, 1,
> and integral n for real x.  Then there's bessel(z,a) which computes
> the Bessel function for complex z, and real order a.
> 
> I propose that we deprecate these names and use bessel_j(z,a) as the
> main routine for all of these.  (What does macsyma call this function?
> Maybe we should use the same name as macsyma.)
> 
> I think this change is safe since the old names are still available,
> and the new name isn't used by maxima code.
> 
> Similar arguments follow for i0, i1, in, with the name besseli.
> 
> Oddly, there's no implementation of y0, y1, etc.  Slatec has routines
> for these as well, so I'll add them with the name bessely.
> 
> There's also g0, g1, gn which I think is the exponentially-scaled
> I(x,a).  Not sure what to call this---maybe an optional parameter to
> besseli?
> 
> I also note that maxima doesn't know any of the mathmetical properties
> of Bessel functions.  A bug.
> 
> Ray
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima
>