3 ms·

> The following design choices may be considered controversial by some: > Dividing by zero yields zero. https://www.hillelwayne.com/post/divide-by-zero/ http

> The following design choices may be considered controversial by some: > Dividing by zero yields zero. https://www.hillelwayne.com/post/divide-by-zero/ https://www.hillelwayne.com/post/divide-by-zero/ That's one interesting point I would love to read any opinions on. Edit: Here's a HN discussion about the `divide by zero` article - https://news.ycombinator.com/item?id=17736046 https://news.ycombinator.com/item?id=17736046

karmakaze 3y ago
There's a brief description of the same choice[0] made by the Pony language. [0] https://tutorial.ponylang.io/gotchas/divide-by-zero.html https://tutorial.ponylang.io/gotchas/divide-by-zero.html
chubot 3y ago
This is notable too: In the year since I wrote this post, Pony added partial division. I still don’t know anything about the language but they’ve been getting grief over this post so I wanted to clear that up. https://tutorial.ponylang.io/expressions/arithmetic.html#partial-and-checked-arithmetic https://tutorial.ponylang.io/expressions/arithmetic.html#par... I think Hillel's post was just saying that 1/0 == 0 doesn't lead to any mathematical contradictions. Another way to think about it is that I believe 1/0 == 42 and 1/0 == 43 are just as valid. They don't lead to contradictions. [In ZF set theory] Since 0 is not in the domain of recip, we know nothing about the value of 1 / 0; it might equal √2, it might equal R, or it might equal anything else. But I don't think you want to use a language where 1/0 == 42, and likewise for 0.
lamontcg 3y ago
FWIW, I tend to write a lot of error checking code that checks for division by zero ahead of time and just returns zero instead. Found an example: public double magnitude { get { double c = max_magnitude; return c > 0 ? Math.Max(c, c * (this / c)._internal_magnitude) : 0; } } I could delete quite a few if statements in pretty hot codepaths if division by zero just returned zero.
- xxpor 3y ago
  This feels like an area where extended/vector instructions could be added to make this fast too. Probably not news to you, but for example, NEON (and likely AVX, I’m just a lot less familiar with it) has saturating addition and subtraction.

karmakaze 3y ago

There's a brief description of the same choice[0] made by the Pony language. [0] https://tutorial.ponylang.io/gotchas/divide-by-zero.html https://tutorial.ponylang.io/gotchas/divide-by-zero.html

chubot 3y ago

This is notable too: In the year since I wrote this post, Pony added partial division. I still don’t know anything about the language but they’ve been getting grief over this post so I wanted to clear that up. https://tutorial.ponylang.io/expressions/arithmetic.html#partial-and-checked-arithmetic https://tutorial.ponylang.io/expressions/arithmetic.html#par... I think Hillel's post was just saying that 1/0 == 0 doesn't lead to any mathematical contradictions. Another way to think about it is that I believe 1/0 == 42 and 1/0 == 43 are just as valid. They don't lead to contradictions. [In ZF set theory] Since 0 is not in the domain of recip, we know nothing about the value of 1 / 0; it might equal √2, it might equal R, or it might equal anything else. But I don't think you want to use a language where 1/0 == 42, and likewise for 0.

lamontcg 3y ago

FWIW, I tend to write a lot of error checking code that checks for division by zero ahead of time and just returns zero instead. Found an example: public double magnitude { get { double c = max_magnitude; return c > 0 ? Math.Max(c, c * (this / c)._internal_magnitude) : 0; } } I could delete quite a few if statements in pretty hot codepaths if division by zero just returned zero.

xxpor 3y ago
This feels like an area where extended/vector instructions could be added to make this fast too. Probably not news to you, but for example, NEON (and likely AVX, I’m just a lot less familiar with it) has saturating addition and subtraction.