Beastly Numbers (1996) [pdf] | Hacker News Reader

Beastly Numbers (1996) [pdf]

people.eecs.berkeley.edu

24 pointsColinWright6 comments

Since people on here really seem to like Unums as an alternative to floats. It is maybe good to point out here that the author of this article has a long series of critiques of them like:https://people.eecs.berkeley.edu/~wkahan/EndErErs.pdf.

This is a postscript file, not a pdf

Tell the mods, they're the ones who changed the title and put the annotation. Title as I submitted:

    Two computers, designed by
    different people, upset by
    the same numbers. (1996)

I thought my title was more descriptive, and it's taken from the abstract of the paper rather than just quoting the paper's title.

Downvoted? Really? I wonder if that was done after the mods also changed to the link to point at the pdf instead of the PS file originally submitted.

There are times I really don't understand people on HN. I don't care about the downvote, because I don't care about the karma, I just don't understand why people would downvote this.

<fx: shrug />

People are weird - I guess I will never understand some of them. But as an engineer at heart, a model-maker and problem solver, it's like sandpaper on the brain when people act in ways that mystify me, and for which I have no effective, working model.

Correct link: https://people.eecs.berkeley.edu/~wkahan/tests/numbeast.pdf

The link I gave works perfectly well, why do you claim that this one is "more correct"? It's the mods who incorrectly added "[pdf]" to the title.

Reply on news.ycombinator.com