Matt Newville

@lellid > I consolidated the duplication of the docstring. OK, thanks. Well, at some point we might fix the capital "I" to "j" (Python's preferred sqrt(-1)). > To the other...

@lellid >what I was really wondering about was why you don't have any tests for the function values in your code? Probably because we feel like we don't need those...

@reneeotten @lellid I'm just going to squash and merge this. If there are more issues to resolve, they can be done afterward.

@Dan-Sk I think this is an argument for making `dill` a required dependency.

@Dan-Sk > It is easy enough to test whether loading with dill works, and to fall back to the other method if it doesn't. I'm not sure I agree with...

@Dan-Sk Well, I'm just not sure that is a matter of "backward compatibility". That is, to even get to this point, you're in the section of the "unpack save file...

@Dan-Sk As I look at this more, I would say, sure, let's add a "try/except" around the "unpack with dill" It might actually be best to reverse the order here,...

@Dan-Sk I sympathize, bu just to be clear, saving a Model is sort of complicated by the fact that it is hard to reliably save a Python function - the...

@Dan-Sk yes, thanks we do understand that- for us it sort of all goes together in the category of "can we save functions in a way that can be reliably...

@reneeotten OK, I basically agree with all of that, at least in principle. My main concern is that it is actually not working well. It seems like the main argument...