Move pywcs' old CHANGELOG into astropy as history.rst - #25

Looks good to me.

Looks fine to me... except that I'm now seeing this error when I build the docs

[u'Wcsprm.dateavg', u':module: astropy.wcs', u'', u'``string``', u'', u'Representative mid-point of the date of observation in ISO format,', u'``yyyy-mm-ddThh:mm:ss``.', u'', u'.. seealso::', u'', u'   `~astropy.wcs.Wcsprm.dateobs`', u'``string``', u'', u'Start of the date of observation in ISO format,', u'``yyyy-mm-ddThh:mm:ss``.', u'', u'.. seealso::', u'', u'   `~astropy.wcs.Wcsprm.dateavg`']:11: (WARNING/2) Explicit markup ends without a blank line; unexpected unindent.

I'm a bit mystified at what made this appear, because I built the docs without any problems after the last pull request (I thought...), and this doesn't seem to have any history-related part. I think this is complaining about there not being a newlline between the first seealso and the next ''string''. But you may as well fix this while you're at it.

I'm not seeing this with current master (1a758b1). I'm on Sphinx 1.0.1, docutils 0.7, numpydoc 1.6.1.

This does point out to me, however, that this still using the Sphinx/docutils ".. seealso::" rather than numpydoc's "See also". I'll make those changes and add to this pull request.

I'm using Sphinx 1.1 and I'm not seeing any errors.

This is not really a big issue, but at the moment the word cd in the history page links to the Python standard library docs, due to the intersphinx extension. Is there a way to avoid this? (just curious as we'll probably encounter this issue elsewhere too)

I think we'll have to fully qualify these using the ~astropy.wcs.Wcsprm.cd syntax. I'll try to find and replace these.

Looks good!

switching to the numpydoc style See Also section fixed the error, so regardless, all is now good.