What are some of your unfinished projects, /anus/?
Let's discuss them to give each other ideas about what not to finish next.
Name:
Anonymous2014-04-13 2:30
I started designing a minimalist, fault tolerant, error resistant instruction set for use as the genetic material in an evolutionary algorithm in which millions programs with only a few thousand instructions competed. It had 18 instructions, four for dealing with partitions in the code that divided routines or whatever it evolved to deal with (findNext, findPrev, findLast, findFirst), two for flow control (jump, jumpEq), five for dealing with memory (move (moves memory pointer), goToBegin (set the memory pointer to zero), getMemPointer, store, load), four arithmetic instructions (add, div, sub, mult), two stack instructions (push, pop), and one instruction to get the current address (getCurrentAddr). Operands were popped off the stack and the result pushed (except for push and pop, which accepted immediate values).
Sadly, I had no idea how to make that fault tolerant such that it wouldn't die if an invalid instruction was executed because the recombination algorithm. I tried reading my SICP, but fell asleep.
Instead I rewrote a dozen or so trivial little algorithms in my assembly language then lost interest.
Name:
Anonymous2014-04-13 5:32
If you keep the same unfinished project for a long enough time, it eventually gets finished.
error resistant instruction set for use as the genetic material 18 instructions
I think you only need four. (maybe name them like A, T, C and G)
Name:
Anonymous2014-04-13 6:20
>>4 If you continuously contribute to the same project over a long period of time, and if the project is realistic in terms of its goals, and you don't have too many other projects that you contribute to, and if you have at least 4 hours free a week, you will eventually finish the project. And by eventually, it will take a scale of a year when you could have finished it working full time for 2 weeks.