This fixes sphinx build errors and warnings in the current version of the docs, adds some docs and a cookbook recipe for OffAxisProjectionPlot, and makes a couple minor adjustments elsewhere.
Since OffAxisProjectionPlot does everything the old write_projection did as well as a lot of extra stuff, I've entirely replaced it in the docs. chummels, is that ok? I'm happy to put write_projection back in if you don't agree.
Update: restoring the write_projection cookbook recipe.
Well, I specifically asked that you keep a reference to write_projection in the docs, not to clobber all reference to write_projection with OffAxisProjectionPlot. But if it is true that write_projection does nothing more than OffAxisProjectionPlot, then not only should be it banished from the documentation, but it should be excised from the code. (We already have too much old duplicate/unused code lying around.)
Frankly, I'm just frustrated that I had a bug directed at me for the last two months to be completed by 2.5 to "fix" write_projection (which was broken by modifications made by Nathan Goldbaum last summer for the Plot Window interface) if all along it was going to be deprecated.
I'd actually prefer that write_projection continue to exist since it can take any image as input. It's functionality is distinct from that in the OAPP. So to me it seems write_projection should be fixed if it is broken (or returned to what it was before PW), and documentation should be included.
Sorry chummels, I guess I misinterpreted your opinion the last time we spoke about it. I've restored write_projection. Sam's point is a good one - perhaps we should call it write_image_with_colorbar or something similar.
I actually don't care at this point. If someone wants to modify it to be some other function name, then be my guest. But I'd just as soon have the docs represent the state of the code, so if it stays, it should be documented, if it isn't documented, then it shouldn't stay. But I'm not going to work on it anymore.