Add NOVAS

[brandon-rhodes]

Now that the Naval Observatory has released a Python wrapper for their NOVAS library, its version of these coordinate transformations (for the ones that it supports) might be usefully included in this benchmark.

http://pypi.python.org/pypi/novas/

[cdeil]

Yes, we should definitely include it. From a quick look it seems that Python NOVAS is

• simple to use and
• has lots of features (e.g. horizontal coordinates, time computations, ephemerides, ...)

[cdeil]

I've added NOVAS to https://github.com/astropy/coordinates-benchmark/blob/master/docs/Tools.rst , but I don't have time to implement the benchmark coordinate computations at the moment.

@brandon-rhodes Is it correct that the Python wrapper is also "Public Domain", like the NOVAS C library?

[cdeil]

@eteq @astrofrog Does anyone know if including a public domain library like NOVAS in the BSD-licensed astropy would be possible? If so it might be a possibility for the base library of astropy.coordinates (see discussion at astropy/astropy-api#6)?
NOVAS looks like a (compared to the other options) small C library that's easy and fast to build (and less complex, but that can of course be debated), and as far as I can see none of the other coordinate libraries have an astropy-compatible license at the moment.

Here in the last section "NOVASPy" in three parts USNO mentions that they do want to add a Pythonic, numpy-aware interface to NOVAS, maybe they would even be interested in doing this in astropy.coordinates?

[eteq]

@cdeil, can you point out where it says that NOVAS is in the public domain? I suppose if there's no license it's that way be default, but if you know of an explicit statement, that would be better. I don't see any reason why we can't use public domain software any way we want, if so.

That said, I'm not sure we want to pull NOVAS into astropy wholesale. But if it's public domain, it's a perfect place to go to get the algorithms we need to build up coordinates, and wrap those as we see fit.

[cdeil]

@eteq At http://pypi.python.org/pypi/novas/ I did find this statement:

This software was produced by the United States Naval Observatory at the
expense of United States taxpayers, and is therefore not suseptible to
copyright, because a copyright would place taxpayer property under
private ownership. Since it is not copyrighted, it cannot be licensed;
it is simply free.

And then I asked here what it means and someone said "public domain".

Caveats:

1. The statement from the PyPI page is not present in the official Python NOVAS release by USNO, I think.
2. I did not check up on if the interpretation "public domain" for this statement is correct.
3. Using Google I couldn't quickly find info if it is OK for a BSD project to include public domain code...

Conclusion: if we are considering using parts of NOVAS for astropy it would be simplest to ask the USNO.

[cdeil]

@eteq Just out of curiosity: Let's assume NOVAS is public domain. Are we then allowed to take some code from it, possibly adapt it, and put it in astropy? Does it have to remain public domain or can / must we put a BSD license on it?

[eteq]

@cdeil - Well, IANL by any means, but my understanding of "public domain" is that you can do anything you want with it without permission. So by that interpretation we are free to pull out any parts convenient and use it however we want.

But you're probably right that we should ask someone at the USNO before we do that.

[mperrin]

You're correct - yes, anyone can replicate, distribute, and create derivative works with no restriction whatsoever from NOVAS, or other works made by US government employees during their official duties. It's probably best when redistributing it to make clear that the portions taken directly from NOVAS are (and remain) public domain works, but that doesn't restrict putting a BSD license on the derivative work as a whole.

IANAL either, but I'm happy to query some of my IP lawyer friends on this one if it's desired. But I'm pretty confident of my interpretation base on prior conversations and some quick reference checks now. See for instance http://www.usa.gov/copyright.shtml and http://www.publicdomainsherpa.com/us-government-works.html

[astrofrog]

I am closing this since the original issue has been resolved.