On Sep 22, 4:02 am, Al Kabaila <akaba...@pcug. org.auwrote:
I wouldn't since I do pretty advanced stuff and I'm happy to use numpy
as-is, but I suspect a lot of people would find a convenient basic
matrix library useful.
Your short list covers most of the important stuff.
Graphics and simulation make use of rigid transformations a lot
(composition of rotations and stuff). It's pretty straightforward and
specialized so I'd give it low priority.
I use linear least squares (psuedoinverse) occasionally for curve-
fitting.
Oftentimes libraries like this have a dumb implementation, that uses
only the basic machinery that ships with every system. You might
consider such an implementation of your API that uses only Python
lists or arrays.
One matrix API that is pretty irritating is Blender's; in fact
whenever I write Blender plugins I usually do all the dirty work with
numpy.
Carl Banks
1. Is there any interest in matrix algebra "for the masses" (I mean interest
in a wrapper for a subset of functions of the packages with a unified
simple syntax)?
in a wrapper for a subset of functions of the packages with a unified
simple syntax)?
as-is, but I suspect a lot of people would find a convenient basic
matrix library useful.
2. What other matrix operations would be required for your area of interest?
Graphics and simulation make use of rigid transformations a lot
(composition of rotations and stuff). It's pretty straightforward and
specialized so I'd give it low priority.
I use linear least squares (psuedoinverse) occasionally for curve-
fitting.
3. What other matrix packages, if any, should one include in the wrapper?
only the basic machinery that ships with every system. You might
consider such an implementation of your API that uses only Python
lists or arrays.
One matrix API that is pretty irritating is Blender's; in fact
whenever I write Blender plugins I usually do all the dirty work with
numpy.
Carl Banks