I find the mechanism in use by Mathematica for naming to
be pretty handy, and you already see it (maybe before Mathematica!)
in macsyma.
That would be to spell out the name and symbol and concatenate them
like BesselJ BesselK
People using %j could do
%j(a,b):=BesselJ(a,b) and/or maybe use an alias renaming, to map %j to BesselJ.
This convention also seems to be used by Openmath, e.g. BesselJ, HankelH1,
but oddly, SphericalBessely not SphericalBesselY.
see
http://www.scl.csd.uwo.ca/~bill/OpenMath/SpecialFns/html/sts/bessel.html
It is clear that just the single letters will not be adequate. Look
at the table of special function symbols in (say) Gradshteyn and Rhyzik
and note that letters like P are way overloaded.
RJF
Raymond Toy (NC/EUS) wrote:
> Sometime ago, we discussed unifying maxima's use of %j and bessel_j (and other Bessel functions) to use bessel_j and friends.
>
> I've taken a look at this finally, and think I know how to do this. Do we want to keep backward compatibility and accept %j? Or just get rid of it completely?
>
> There are a couple of other related issues. The docs for specint show that there's also %p (Legendre functions?) and %m (Whittaker's confluent hypergeometric function?). Do we want to replace these as well. I don't think maxima knows anything about %m. Maybe the specfun package knows about %p. (Sorry for the vagueness. I don't have the docs with me right now.)
>
> Not to mention hstruve[n](x), li[n](x) and so on. These should, perhaps, be changed to hstruve(n,x) and li(n,x). Or something better for li.
>
> Ray
>
> _______________________________________________
> Maxima mailing list
> Maxima@www.math.utexas.edu
> http://www.math.utexas.edu/mailman/listinfo/maxima