I wonder if you could convert it to colour as well, if you changed the rule from a single thread for the whole thing to three threads for the whole thing (CMY). Might be quite an interesting effect since you'd need to sort of "smudge" colours across the image the same way you're essentially smudging the blacks.
Edit: Wait a minute, actually it wouldn't work because you can't combine colours. Put magenta thread on top of cyan and you'd get... magenta.
I wonder how much better a less greedy algorithm would be-- e.g. one that considered many possible moves a few positions deep, and took one step towards the best... or one that after finishing unwrapped the beginning halfway and rewrapped it backwards from the middle, then unwrapped the end, and rewrapped it forwards from the middle, and repeated until convergence.
I'd guess that the calculations are essentially a reverse radon transformation. Each string should be analogous to the sensor reading for a cat scan of a density field matching the portrait.
I didn't see it mentioned, but I assume you have to do it with a single unbroken string, so it's a bit more involved maybe. I'd imagine you'd first generate all the lines that best approximate the image, then insert additional lines to turn it into one unbroken string where they're least disruptive.
> Over 2 billion calculations are needed to produce each pattern; not much of a load for today’s computers, but definitely an impossible task for the human brain. So, this is a new and unique type of knitting that could not have been implemented a few decades ago, without computers.
It's hard to take this assertion at face value. Compare that to what a human go player can do vs. a computer. (2 billion computations would be on the order of seconds on a normal computer)
I wouldn't comment except that the OP's comment history here is also highly critical, so: they already posted this like 8 months ago. It's also not knitting as most people understand the term. I mean, it's kinda neat, but the title here is misleading, and it's the second time they've used it. Promotional?
By no means! We invited gus_massa to repost it. This is part of our ongoing experiments in recovering good stories that fell through the cracks the first time round—described at https://news.ycombinator.com/item?id=11662380 and many other posts linked from there.
Also, it's fine to repost a story if it's good for HN and hasn't had attention yet. Otherwise too many great submissions would languish unseen. This is important enough to be in HN's FAQ: https://news.ycombinator.com/newsfaq.html.
Putting [repost] would probably have an impact on the vote ratio up and down given that in many online communities 'repost' is seen as a negative thing perhaps.
If you can suggest a more accurate and neutral title we'd be happy to change it. It doesn't seem misleading to me at all, but I know nothing about knitting.
http://www.thevelop.nl/blog/2016-12-25/ThreadTone/