Hacker News Reader

Comment by gerdesj | Hacker News Reader

Blimey, I just used a spreadsheet 8) I'm not sure what language you are using but it actually comes close to looking like a proper proof - all the inputs are on show and the construction is reasonably obvious. My spreadsheet is trivially correct as well but some of the working is not on obvious show unless you inspect the cells. Also, my spreadsheet is unavailable unless I share it.

Nice one.

View on HN

Be careful accepting JS numerical proofs, since they're all doubles. e.g.

    > 12345678910111210 + 1 == 12345678910111211
    true
    > 12345678910111210 + 1 == 12345678910111212
    true

My favorite one is disproving Fermat's last theorem:

    > 19774**3 + 257049**3 == 257088**3
    true

> Be careful ... e.g.

> two true statements

What?

The two 'true' statements show that 12345678910111211 == 12345678910111212 transitively, which is obviously not possible.

Look closely. Or, equivalently

  > n=12345678910111210
  > n+1 === n+2
  true

It's JavaScript, ES2015

Thanks. Google ECMAScript to me is the odd thing I had to learn (I have a really odd accent when speaking it) to get Novell's DirXML working.

Reply on news.ycombinator.com