Re: definite integral of arbitrary function

[rdurkacz]

With this and other help from the mailing list I bellieve I have passed the
getting-started stage with Octave. I have an observation:

It seems that vectorisation is an important part of Octave. Going by the
manual, the intention is that functions are written to support
vectorisation. Basic as they are, the quadxx functions do not support it.
It is pointed out that we can vectorise quad using arrayfun, but presumably
that is inefficient in execution and in the amount of code-writing you have
to do.

lsode can do quadrature as a special case and is vectorised. It has various
special methods built into it, whereas the quad functions come in several
different methods, each a separate function.

It seems there is room to harmonise the quad functions with lsode to provide
vectorisation and pick up other benefits. The fortran code for lsode,
together with a c version from the Euler project, is interesting... (I have
not looked at quadxx code.)

I am interested in what Octave users think about this.

--
View this message in context: 
http://octave.1599824.n4.nabble.com/definite-integral-of-arbitrary-function-tp4578474p4613833.html
Sent from the Octave - General mailing list archive at Nabble.com.