Re: Request for comments: imread and imwrite

[Michael D Godfrey]

On 10/08/2010 06:37 AM, John Swensen wrote:

I know that if I don't match behavior exactly, there will inevitably be some code out there written in Matlab that relies on the dimensions of the colormap after reading in a file to run correctly (or something akin this scenario) and there will be a question (or criticism) about the code not working for Octave.  So in some sense I would like to make it completely compatible but am dreading the process.

John Swensen

Matlab compatibility is a complex issue and, obviously, there is no
right answer for all cases.  Here are a few guidelines that I would suggest:

1. The quality and efficiency  of Octave are essential (i.e. they dominate compatibility).

2. If exact compatibility can be achieved at low cost it should be done.

3. Errors in Matlab should not be reproduced in Octave regardless of
    compatibility.  Deciding what is an error is itself a problem in many cases.

It might help if there were a definite "policy statement" about the Octave
view on compatibility with Matlab and with other products (such as R?).
The motivation for compatibility should be, I think,  to to be most helpful to
users while providing the highest quality product.

As a separate thread I am sending an important Note by Kahan on Matlab.


Michael