It seems to me that an often requested feature is to be able to have POV-Ray
smoothly "plot" x,y,z data in some way.

So, I was wondering how much development would be required to:

- Directly generate a grayscale image / pigment / pattern / function or
heightfield using array values, CSV files, or other easily user-generated data
source.

- import text data to generate a df3 density function

It just seems to me that when these binary file types are imported, that data
gets imported into some data structure, and it seems a short jump take data in
an existing 2D or 3D array and copy it over to an image_map or df3.

- read and write BINARY data directly from and to files, facilitating the
creation of required file types without the need to install software, acquire
proficiency in another language, it's syntax, etc.

Also of use here would be determining the type/cast of data, such that data
could be read from a file without prior knowledge of the exact file structure.
I believe there's been a related concern about determining if an identifier is a
vector or scalar.

Some sort of error handling would be extremely useful as well, such that
erroneous data could be rejected and just skipped instead of flat-out failing
the render, and this would be useful as a way to reset the file pointer and try
to read the correct data type (as a workaround).  One could also skip the usual
degenerate cylinders and cones, etc if desired, to simplify certain coding of
scenes.
And of course - the cryptic nature of the error messages is sometimes about as
illuminating as "Error 512".

Post a reply to this message