Store angular measurements consistently.

[timstaley]

(CF https://redmine.astron.nl/lofar_issuetracker/issues/4742)
As of 2.0, the database schema uses a mixture of arcseconds and degrees, which requires extra care over conversions. At the risk of invoking numerical precision issues (which we should perhaps worry about a bit), it would be nice to store everything consistently. The simple option would be degrees everywhere, but John pointed out that radians would be more efficient when it comes to trig functions. (Though my worry there is that when we're doing tricky meridian-shifting corrections, working modulo an irrational number might cause precision issues of its own).

[AntoniaR]

Apart from making it more consistent, what are the advantages of this? This issue would also mean changes to Banana and careful checking through the documentation to change all the units. It sounds like it could be a lot of work to implement (a lot of extra care would be needed to check the replacement of existing conversions through out the entire code) and you already raise the loss of precision issue. The database is already well documented and it should be very clear what units each column have.

[timstaley]

I'd argue that consistency is something that's pretty valuable in a codebase. But you're right that it may be a costly bit of refactoring for not that much gain - I'd still like to leave it here for evaluation when there's time.

[AntoniaR]

Sounds good.

[AntoniaR]

We should consider if this is important/useful for R7