>>14,16
I've now corrected everything you've mentioned but splitting out the functions from naive-append-vector!, removing the hard coding, and the inefficiencies in copy-append! there (you are completely correct, and of course it matters more than SIMD being available). I'm not a particularly experienced programmer I'm afraid to say, and as you've no doubt uncovered that from looking at my library here. I tried to look over your spiral code to help you a little since you've helped me so much but I've never done anything with graphics, python, or objects, and I'm afraid I couldn't come up with anything which might be of use. I do really appreciate everything you've done for me, and maybe one day when I have a bit broader set of knowledge or when you're working in my wheel house I can give some sort of a valuable suggestion. Anyway don't feel obligated to continue helping me, you've done so much for me already and I completely understand if there is some other better use of your time, if you are going to continue or if you just want to see the impact of your changes thus far here's a updated version: http://ix.io/1wVh
>>15
Those sound like pretty neat projects anon. I actually started on a image archiver for imageboards in guile awhile back, but it turns out that guile (& gnuTLS) don't play well with openbsd so I had to abandon it before I could even give it a test compile, it's not much (pretty rough if anything) but it might be of some use to you: http://ix.io/1wVg If you have any questions or you need help I or someone else here might be able to assist you.
I finished The Unix Hater's Handbook last night. I used to be really hooked on the New Jersey stuff, but these days I'm really fond of "The right thing". It's nice to be able to laugh at how far you've come. Also I think I will go ahead and try to optimise my radix tree a bit more, now that with the help of anon I've abstracted out my extract-5 function in favor of a function that simply returns the correct position in the current vector I can change the underlining function more easily, making it seem more reasonable to implement the relaxed variant of the radix tree (a sort of halfway house between the adaptive and the general) https://infoscience.epfl.ch/record/169879/files/RMTrees.pdf