3 ms·
I thought it would be a fun thing to do tonight until I saw all the tips given at the end of the article. They basically give the solution, it's not funny anymo
by uniclaude 15y ago
I thought it would be a fun thing to do tonight until I saw all the tips given at the end of the article. They basically give the solution, it's not funny anymore :(.
- stevelosh 15y agoSo make it more fun: * Support crosscut shredders (shredded in both directions). * Make it work even if the shreds aren't all the same size. * Make it work when there are some pieces missing. * Show where those missing pieces would probably go. * Make it work when two images' shreds are mixed together. * (with missing pieces) * (generalized to N images) * Predict what the missing pieces might look like (this one is much tougher). * Think of ways you could cheat (like searching Tineye for the fragments to find the original image, if it was originally online somewhere). You can keep going as long as you want. If it's too easy, just make it harder. Edited because HN doesn't support Markdown or linebreaks. :(
- bfung 15y ago* write a plugin for photoshop or gimp that will select the arbitrary shreds. Does that exist already?
- emmelaich 15y agoAnother idea: combine it with a file recovery or computer forensic program to stitch jpeg file fragments together.
- mnutt 15y agoAlso: make it work for shreds at any angle and overlapping, as if you dumped the pieces on a flatbed scanner and hit 'scan'.
- gerggerg 15y agoImplement zoom and enhance so we can see what all the people in those buildings are up to.
- uniclaude 15y agoIf it's too easy, just make it harder. Absolutely. Wa can make it harder in many creative ways, and that would be fun, but I wrote this comment to express the feeling of disappointment I had when I saw the post. I felt a little bit like someone showed me a candy but already ate half of it.
- Terretta 15y agoAgreed. Showing how to use PIL, fine. But if they really intended to use this as a gate, then showing the key to a solution -- what's the point?
- patio11 15y agoIt is still Fizzbuzz++: they're not looking for computer vision researchers. They're looking for programmers who are competent enough to be worth their time interviewing. If you can go from hints in the direction of the solution to a) reading library documentation, b) coming up with an innovative solution to an only-partially specified problem, and c) actually producing functioning code which implements your solution, you are heads and tails above the hiring pool.
- uniclaude 15y agoI have a serious problem with this. I agree with you, and I know that you have some serious knowledge about that, but seriously, reading HN makes me feel that this hiring pool would be filled only with programmers that could solve any type of code-puzzle in their sleep.
- tptacek 15y agoThis particular problem isn't a puzzle. It really is a Fizzbuzz for image processing. No qualified applicant will fail to implement this.
- dabent 15y agoI have to echo patio11 below. The point here isn't to find PhD-level image analysts, it's to find some relatively solid (Python) developers in fairly short order, while handing out advertising to people in the form of shirts. The problem isn't hard for people who end up reading Hacker News on a Friday night (and Veteran's Day in the US, when some folks are off work altogether), but this is just the kind of people they're looking for in a tight tech labor market. I noticed they did remove on of the functions they had listed in the hints at the bottom, possibly to make things a bit more challenging.