[Date Index] [Thread Index] [Author Index]

Interpolation

• To: mathgroup at yoda.physics.unc.edu
• Subject: Interpolation
• From: weller at vuse.vanderbilt.edu
• Date: Tue, 26 Jan 93 12:00:54 CST

I have an observation and a question.

1) Observation: It would be really nice if the function Interpolation
also took an alternative argument structure similar to Plot and then
used the same adaptive algorithm that Plot uses to find automatically
a good set of points to use.  I know that one could hack this by plotting
the function you want to approximate and then extracting the data from
the graphical object (it's [[1,1,1,1]] isn't it?) but this is not elegant.
A possible enhancement for 2.2? Are you out there WRI??

2) Question: Does anyone have the code that Plot uses to build up its
list of plot points?  If not, would WRI be willing to make it available as
a function which returns the list of plot points?

By the way, the function which I am approximating is defined by a numerical
integral. It's a slowly varying function of its argument but it takes 
a long time to compute each value. Also it isn't usally obvious where
to place the greatest density of values.

Robert A. Weller
Vanderbilt University