Notation about acefit.jl

[cortner]

I like c or theta for parameters, y or Y for observations, and I don’t mind A for design matrix.

There is a potential advantage of using capital then y could be a single observation or group of observations associated to one structure

[cortner]

x is usually a State or other input variable

[wcwitt]

Various views from Slack

"If we do unicode, any python interface will probably just have to use Phi instead of actual unicode, but I'm OK with that."

"Python is actually fine with unicode variable names now, just not infix operator. you can even type \Sigma in ipython and jupyter just like in Julia"

"So it's not a technical hurdle, but I'm still not a fan of making the typing more complex. Too much like APL"

"Also (and I suspect I'm in a tiny minority), I'd actually prefer a more verbose notation (even with its attendant typing load). I'm still entirely unable to remember, in nested sampling, the meaning of N vs. K."

[cortner]

I'm ok to enforce no unicode in library code and only allow it in scripts.

[wcwitt]

For now it's AC=Y. Closing.