Variable namespaces and naming conventions

[whoburg]

Should I be able to use V for velocity in one model, and V for voltage in another? How do we combine these models? How do we keep track of variable naming in a cross-organizational optimization model?

[bqpd]

@whoburg will be making an aircraft example with a test system to which increasingly complex models are added, which may help us consider this issue.

[bqpd]

merging #65 into this and closing that one

[bqpd]

@whoburg:
"
For example, in a drag model, should I create a variable D_fuselage? Or should I just put the (monomial) fuselage drag expression inside a larger posynomial? Do we offer a way for users to name terms in a posynomial? What is our recommendation for new users on best practices?

@bqpd says this is related to variable namespaces and locality.
"

[bqpd]

do we want to create a special "objective" variable, or just encourage the user to put a single variable (perhaps inverted) as their objective function?

[bqpd]

1. There's now a cost element of the GP solution.
2. as variables are now VarKey objects, it was simple to implement a model attribute which distinguishes two variables, giving us namespace separation.

The main thing we're waiting for is an example importable model, now.