Lisp jarvis march

[Trashtalk217]

It works, but that is mostly all that can be said about it. I am nowhere near happy with it, but I want people to review it and give some tips. That's also the reason I haven't editted the .md files for Jarvis March. I'll be working on it way more, but for now you can enjoy the spaghetti.

[Trashtalk217]

List of things I'm still not happy with:

• The way I currently calculate angle is nice and compleet, but I'd like to cut (some) corners.

• The return-from in line 106 works, but I should just write the loop properly (loop until and then return) instead of jumping out of it like this.

• Could maybe make the pointstruct more convenient (read-only, shorter way to read values, etc). Both leftmost-point and next-point-on-hull are doing something really similar, so that could be abstracted, I'm also getting tired of tail-recursion, but this might have to do. => abstracted the two into extreme-of

• The p-eql is poorly named and maybe I could somehow integrate it into the point struct, or another comparison function can be used (there are 4 in common lisp!) => Replaced with equalp

The reason why I removed the WIP label, is because I want additional input, besides my own. I expect that there'll be more people looking to review, when it doesn't have this label.

[c252]

Just some small things I noticed

[Trashtalk217]

I've tried, but I don't know how to shorten the functions involved in calculating the angle. My linear algebra is a bit rusty and the way others do it looks (and maybe is) quite different. If anyone can elaborate that would be very much appreciated.

[Trashtalk217]

I've fixed all the things I was unhappy with. All I need now is a review!

[jiegillet]

It looks good, I can understand everything fairly well. I am not a fan of loops and I'm sure there are things that could be refined but it would probably be matter of taste at that point.

[jiegillet]

Let me know if I missed something.

[jiegillet]

I think I realized you don't need second-point-on-hull. Alls it does is call next-point-on-hull you might as well do that in your loop.
You can start your hull with the (- (point-y start) 1) and start points. And then remove the fake point later.

[Trashtalk217]

Are you sure this is a good idea, because the the first element of the hull, as defined in the hull list, will then become the starting point, so the check after that until (equalp (first hull) start) will pass and the loop will end, I cannot find an elegant way around that, do you have an idea?

[Trashtalk217]

I could place an extra check like if length hull is greater then 2, that might work.

[Trashtalk217]

Honestly don't know if this is an improvement, but it's a couple lines less once more.

[jiegillet]

I think it's an improvement, removing one function help reading the code.
One last thing, I stumbled onto this could you use it to get rid of [and] (> (length hull) 2))?

I apologize if my ideal piece of code is the shortest that can be. Feel free to fight back especially since I don't actually know lisp.

[Trashtalk217]

I wouldn't have expected anything else from a haskell programmer, @jiegillet. But, jokes aside, I mostly agree with you. There's something very satisfying about getting your code under 100 lines and if you're not cramping everything together it more often then not makes it more readable.

I, however, do not understand what you mean in you're request. Could you please be a little more clear?

[jiegillet]

Something like:

 (loop with start = (leftmost-point gift) with hull = (list start (make-point (point-x start) (- (point-y start) 1))) do (setq hull (cons (next-point-on-hull (first hull) (second hull) gift) hull)) while (/= (first hull) start) 

(although I'm pretty sure my syntax is wrong at more than one place)

[Trashtalk217]

I actually can't believe it worked, but it did.

[jiegillet]

I must be a Lisp genius! Imagine if I actually learned it ;)
One last copy/paste incident to fix and we're good to go

[jiegillet]

You imported it twice!

[Trashtalk217]

It's finally done I think, thanks for reviewing @jiegillet and @c252