Make handling of MEF files easier

[mwcraig]

Though an example of working with MEF files was added to the documentation in #694, we should make it easier to work with these files given the growing number of mosaicked cameras.

[MSeifert04]

One of the main problems I had with MEF and CCDData was that data/mask/uncertainty produce and aweful lot of extensions. Also the files can become enormous compared to the raw data - at least twice as large because uncertainty is as big as the data, often it's even more than 6-8 times as large as the original file.

Apart from that it's also a bit difficult to implement correctly, especially if memory could become a concern.

But I totally agree that having some MEF-handling convenience functionality could make it a lot more accessible. 👍

[crawfordsm]

There's also an issue of inconsistent usage of how to indicate data vs. mask vs uncertainty in the different extensions, and how the different extensions map back to detector space (usually split by amplifiers for example, but could also not be split and just different extensions).

There are a number of telescopes that have also implemented slightly different solutions as well (Gemini, MMT, JWST), so we might want to do some research and look at what different groups are doing.

Generally though, I'm +1 as well to adding it