A sunday reflection

The guts of the manim library seem to lend itself well to things of a functional and topological nature. In some ways this makes sense, we want to be able to say things in the programming (functional) and it do things on the screen (topological). I wonder if that creates something like a skew in the kind of stuff you make as far as topics are concerned. Which begs the question, for me at least, if I wanted to do something neat in logic (model theory) how can I express it in a more visual way? I wonder if I could do stuff with homotopy type theory (HoTT), that might be kind of neat.

I think with the chicken breast stuff that I need to work from beginning to end now. I feel like I have a rough idea of how it is spaced out as a problem, and now I just need to go ahead and tighten the topics up in order. The two big things this project is going to deal with is cell differentiation, and bioreactor stuff. Once I know what cells I am going to deal with, then I can worry about the constraints of the bioreactor. The bioreactor stuff is going to be somewhat harder than the cell differentiation stuff, because this is where the rubber meets the road. I am thinking about having a vascular system, nerve system, and musculature all working together. At this point I feel like a kid listing things off to Santa Claus. The creation and growth of each tissue type is understood, but I am not sure about the integration between any two let alone all three. Push come to shove, I will have to look at the theoretical and iterate back and forth with simulations to create a decent hypothesis for a prototype. While this may take longer, I believe I would understand the topic better and it would result in less waste by blindly creating experiments or prototypes.

I'm thinking on doing some stuff on an algorithm I think I created. It uses the Wasserstein distance in an iterative fashion, and has potential applications in a variety of fields. I came up with it over a year ago, and I think it is time for it to have a home. It might be kind of neat to do a visual of it using manim, the algorithm (Wasserstein Walk) deals well to visual stuff I think. If I do, I'm thinking about sharing the link with places like HackerNews maybe?

I wonder how I could go about simulating biological cells. What kind of assumptions would go into various modeling styles. Something like LAMMPS, we would have n-body problems (n particles have n^2 interactions). But I wonder if I could take something done in LAMMPS, document the operating assumptions, and describe what is happening equationally. That way I could just check what's going on, is assumptions invalidated, if not the use set equation for set interaction. So, it would be like a macro-micro simulation, but we are using the equations to abstract away the micro simulation. It would be more like macro-equation/equation-micro kind of simulation, maybe?