[AstroPy] Deployment and packaging

[Tommy Grav]

On Jun 14, 2011, at 10:11 PM, Perry Greenfield wrote:

> 
> On Jun 14, 2011, at 9:44 PM, Tommy Grav wrote:
> 
>> 
>> On Jun 14, 2011, at 9:12 PM, Perry Greenfield wrote:
>> 
>>>> - Base Python binary installation with NumPy, SciPy, matplotlib
>>> 
>>> But how to distribute binaries? And what about non-Python tools that
>>> are useful in the suite of astronomical tools?
>> 
>> Maybe I am misunderstanding, but when I am looking for a python
>> astronomy package I don't want a slew of non-python tools. What
>> I am looking for is packages that are useful as packages to use
>> in my own python codes. pyfits and pywcs are great examples of
>> clean python packages that are fairly easy to install and use
>> in python code. I am hoping and advocating that astropy/astrolib
>> becomes something like scipy, rather than just a bag of disparate
>> package that have little on common than astronomy.
>> 
> You may not, but many others do.

No doubt. I can only voice my opinion :)

> IRAF is only one example. SExtractor is another. And there are others. There isn't yet enough to replace these, and won't be for some years (if ever, it is probably silly to think everything we need will be in Python, or that there won't be some very useful things out there not written in Python).

My opinion is that those tools belong in as separate packages. To me they just don't
belong in a python astronomy. I guess what I want is a astronomy library, rather than
a package collection. I guess that the SExtractor could be included as a C library with
a python front end, but IRAF is just a pain to install and to me constitutes a rather
hard hill to climb. Also my hope is that a proper astronomy library with C extension finally
someday make IRAF absolutely obsolete. 

> Note that scipy is not easy to install everywhere.
> 
> So, certainly, standard Python installs for separate tools should be available. We aren't talking about making that go away.

I understand that. My point, however, was that the cramming of everything astronomy into these
tool-collections is in my opinion what has been the problem for the current libraries/packages 
from becoming the defacto standard as scipy and numpy has. Your milage may vary. 

Cheers
  Tommy