Ask HN: What has binary to do with computers? | Hacker News Reader

Ask HN: What has binary to do with computers?

3 pointsColinWright7 comments

An intelligent, but not very tech savvy, person asked me today: "What has binary to do with computers?"

We didn't have time for a really long, complete, detailed, start-from-the-ground-up answer.

So what would you have said?

I would have probably said (with some imprecision): Electrical circuits either carry a current (binary 1) or they don't (binary 0). [Ignoring the fact that you could also measure the AMOUNT of current carried here] Because of this underlying limitation in electrics you need to make do with just a binary system in computers.

Depending on the amount of time remaining I would either go into more depth or point him towards "Code: The Hidden Language of Computer Hardware and Software" by Charles Petzold (http://amzn.com/0735611319)

I read that book as a child and loved it, but it is so hard to find a book named "Code" when that is all you remember. That link just made my day.

I would explain that it's very difficult to store and measure precise charges, which is why computers are digital (on or off). Therefore all arithmetic and logic functions are done in binary numbers, as you can only work with on and off switches. The underlying hardware itself only allows this, until something better comes along.

Because binary allows for exact copies.

Imagine copying a cassette tape or VHS and then copying from the copy. Eventually you have noise. Same for photocopies, and obviously all of these are somewhat dated.

This property of noise rejection gives digital computers a significant advantage over analog computers.

Everything - computers are electrical, essentially billions of on-off switches. On and off is best represented by binary, so at the core of how a computer works is binary (on-off) math and logic.

ACP, Knuth: SETAN(rus) balanced, unbalanced(balun) ternary, (0,1) arithmetic encoding, Bechet's weights with Euler's proof... with Nicolai's dual(fewer Ohaus weights with greater nos comparisons conserved) solution, and remind savvy person of analog integraph, and instant immediacy of K+E slip stick.

fyi: o(0,1) >> o[0,1]

Reply on news.ycombinator.com