Guiding principles and first steps

[jessecusack]

Following our meeting this morning, here are our goals for this package:

• Produce the highest quality possible estimates of epsilon and chi (with error estimates and georeferencing)
• Be very well documented and transparent (lots of inline comments)
• Be easy to automate (fit into ship based operations where new data comes in bit by bit) and configurable using parameters
• Following the spirit of ATOMIX standards, e.g. L1, L2, L3, L4 processing and outputting standardized data that is CF compliant

Important, but secondary goals:

• BBL focused processing
• Handling of DO/Ch/Turb data

Things to think about down the road

• Calculating chi. FP07 calibration drift?

Code style

I don't have a strong preference for camel case or snake case, but I want us to be consistent. We may want to consider following matlab community guidelines: https://sites.google.com/site/matlabstyleguidelines/naming-conventions/functions

We may want to consider appending some tag to our function names, like m_map and gsw do (e.g. m_plot, gsw_Nsquared) to easily identify our functions and avoid namespace conflicts.

[mousebrains]

I'm slowly switching to snake case, and we can migrate naming conventions to be roughly in agreement with the "community guidelines" Jesse linked to.

For a prefix, I have been using osgl_, Oregon State Glider Lab, for my other public packages. We can use that, or come up with something else. But yes, it would be a good idea for code isolation.

[jessecusack]

I don't like osgl, but happy to brainstorm other options.