>  On page 15 in the PIL handbook, the following is said about the
>  Image.new function:
>  
>     new( mode, size ) or new( mode, size, colour ) creates a new image
>     with the given mode and size. Size is given as a 2-tuple. If colour is
>     not given, the image is initialised to solid black.
>  
>  The last sentence is not true; if the colour is not given, the image
>  is not initialised at all.
>  
>  Now, the only problem is whether I should change the documentation or
>  the implementation.  Opinions?

I assume that initializing the image takes a reasonable amount of
time.  It is useful to have a way to avoid wasting this time (often
while the user is waiting!) for applications that know they are going
to set every pixel anyway.

So unless you are prepared to provide a different interface for
creating an uninitialized image, I'd say, fix the documentation.

--Guido van Rossum (home page: http://www.python.org/~guido/)

_______________
IMAGE-SIG - SIG on Image Processing with Python

send messages to: image-sig@[...].org
administrivia to: image-sig-request@[...].org
_______________