That’s nice to see this on hn. Some might recall a website called flipcode, the author used to write articles there under his handle “phantom”. Ompf is a community for raytracing enthusiasts, where real time raytracing was pioneered way before the adoption by e.g. gpu manufacturers.
Wow, I do remember flipcode (and phantom). Strange to see those names after so many years. It was from a different era of the web when it was full of niche, useful sites like these populated by domain experts. I miss those days :(
Twitter is a good source of information now, and it in fact looks like the free tutorial is making a comeback with Twitter as a podium.
- J. Bikker aka Phantom. :)
Probability of a hacker news top 10 breaking your hosting: high
Interesting though, I don't think I've ever seen a code 507 in the wild!
Does it mean that the logs got so big that the httpd wouldn't serve requests, for want of space to store the record?
I don't know actually. 507 was introduced in rfc 4918 alongside WebDAV, so my naive assumption is that its main use case was for something more associated with authoring or general file manipulation.
I'm surprised that it has much use for a normal GET request, but I guess anything can be a bottleneck!
I vaguely remember a good probability mathematician talking about an optimization as possible with sorting what objects to render and it being much more efficient during my time in IT in the university. Ie with math you could exclude a lot of the objects to render. Happy to see it implemented and looking forward to game with it!
I only glanced through the article and wonder if it would be easier to use deterministic grid or adaptive grid?
...you should probably read the article since it covers that.
Ok, I read the article. I guess this is more educational exploration rather than practical advice. So conclusion is that deterministic grid or adaptive deterministic grid (not directly discussed) beats random sampling.