[SciPy-user] io.read_array in SciPy vs the one in "scientific"

[Travis E. Oliphant]

Fernando Perez wrote:
> Sebastian Haase wrote:
> 
>> Hi,
>> Quick question: How does SciPy's read_array compare to 
>> Scientific.IO.ArrayIO
>> (cf. http://starship.python.net/~hinsen/ScientificPython/) ?
>> And why are there these two (independent !?) projects anyway ...?
> 

For one, scipy is larger than ScientificPython by a large margin (a much 
larger code base).  The naming conventions are different.  We would 
gladly include code from ScientificPython (and have done so in fact) but 
need to change the naming conventions.

io.read_array is based somewhat on ArrayIO but goes beyond it in terms 
of what is supported.   SciPy is trying to gather code together in one 
super-package.  Code that could be adapted from ScientificPython is 
always welcome.

> 
> Well, for one, the Scientific readArray allows negative numbers, so it's 
> an improvement :)

A bug that was fixed over a month ago.

> 
> However, the issue of the two projects is there.  After my little run-in 
> with the negative numbers issue, I turned over to Scientific as plan B.  
> I view SciPy as the 'lead' scientific computing project for python, but 
> I wish there were a bit more code sharing, especially in areas where 
> Scientific already paved the way (such as the readarray case).  But 
> there may be licensing issues I'm not aware of, so please don't read too 
> much into my opinion on this matter.

read_array did borrow from Scientific.  It's just that significant 
changes had to be made in order to support the full flexibility desired 
from read_array.  io.read_array is quite a bit more flexible.

We have not ignored ScientificPython but borrow from in when possible. 
Contributions are always welcome.

-Travis O.