[fsloth]

I would prefer people read the original paper than any second-hand explanation of it. The paper is very readabale and understandable, and is probably relevant as long as into the future as human beings write code.

[majewsky]

When mcnichol said "make this its own article", they probably meant to submit the link to HN as its own submission, not to blog about it.

[_57jb]

Yeah, that's what was meant. Totally agreed with the statement above this one, it reads well.