typochon @typochon@functional.cafe

"This problem is interesting for modelling large systems, cyber-physical systems, IoT, insert your favourite buzzword here..."

Quality presentation ^^

@amiloradovsky I don't think so, it was something someone made using those (I assume). But it was an example of what was possible rather than a theory

I checked, just in case, but didn't find anything

Some time ago, I saw a toot with a link to a #Coq proof involving real numbers, but I can't find it anymore :(

I think it was about proving that the ratio of two Fibonacci numbers converges to the golden ratio, but I'm not sure.

Anyone remembers it and knows where I can find it?

Everybody's talk: loads of math and equations

My talk:

meme

@otini The small-angle approximation is a lie

I wanted to explain the code you get in Haskell to write this sort of model, but it's actually quite tricky to explain in a toot^^

Hopefully I'll write a blog post about this when this is more polished

Forgot the units, it's radian for the angle and radian per second for the angle's variation

On the previous graph, the curves kind of looked like sine wave because the angles remain small. But, this isn't true anymore for larger angle, as illustrated by the result one gets if the pendulum is released from 3π/4 radian (135 ° from the vertical)

So now I can make graphs like this and write my equations in #Haskell!

This is the result of a simulation of the movement of a frictionless pendulum, suspended to a 1 m rod for 10 s. It starts with no velocity at an angle of π/6 with the vertical. The solver returns every 0.1 second. Blue curve is the angle over time, orange curve is the derivative of the angle over time.

It's not perfect, for instance, I still have one weird problem with the types of some function pointers not being imported correctly. But, while I didn't test anything until I had done everything, I haven't had any segfault or unexpected result with the generated bindings, and I find that pretty impressive!

So after a few days at working with #c2hs to make bindings to #sundials in #haskell, I'm actually impressed by it.

It's a tool to help with writing bindings to C library. It can generate a lot of the boilerplate associated with this sort of stuff (e.g., generating wrapper types on the Haskell side) and typechecks the FFI imports by analyzing the C types in the library's headers.

I feel you, Sundials… I feel you

Me trying to understand the ACM copyright transfer page

Everything I've tried has failed, either because I can't use void* as the type of some hooks (either it can't parse it, or it complains it's not declared anywhere or it just fails with an internal error, depending on the hook I try to use).

Does anyone know if there's a way around the problem, or if I'm just condemned to use Ptr () for this?

I'm trying to use #c2hs to generate bindings to #sundials.

Sundials has a function IDACreate, wish allocates a block of memory to then pass around. This block of memory is simply of type void*

When I import this function with c2hs, I would like to make the Haskell version return a newtype wrapper around the pointer, but c2hs doesn't let me do that.