Re: A question on Matlab compatibility

[David Bateman]

Peter Jensen wrote:
> Gents,
> 
> In relation to the discussion below,
> could somebody clarify the procedure 
> for reporting differences between mat*ab 
> and octave-forge.
> 
> I found out that the scaling for pwelch in 
> octave-forge is different from matlab,
> but I was not sure  if I should report this 
> as a bug.

Well, I suppose the proper place to report it would be
address@hidden as this is an octave-forge function.
However, anyone who is likely to help you fix is is subscribed to both
lists and so it probably doesn't matter

> If the scaling of pwelch was changed in
> octave-forge it may break some peoples 
> code. On the other hand the function is
> not mat*ab compatible right now.

The consensus recently has been that the pain of breaking octave's own
interface with itself is the lesser of the pains than dealing with
incompatibilities with matlab, and so changes to bring the API into line
with matlab will almost certainly be accepted both in octave and
octave-forge.

D.