2

There are many ways to represent data using binary like - unsigned, signed magnitude, 1s/2s complement, offset-M, floating-point, ASCII, and Unicode. Why do we need so many different ways?

Peter Cordes
  • 328,167
  • 45
  • 605
  • 847
sunny
  • 149
  • 8
  • is the data that you are representing all of one type? – jsotola Sep 11 '21 at 16:11
  • Representations are *interpretations*. Why we have many interpretation of the same number? Because we have different problems. – Margaret Bloom Sep 11 '21 at 16:35
  • We don't. 1's complement and sign/magnitude integers are not used in modern systems, except maybe in some file formats. And Unicode is a superset of ASCII, so if you decide to use Unicode everywhere, you don't need ASCII anymore. (Although then all your functions have to handle the general case of Unicode where a character isn't fixed length.) – Peter Cordes Sep 11 '21 at 21:50
  • It is useful to have efficient fixed-width types, and variable-length arbitrary-precision formats, though. And to have a text character set so we can print integers into human-readable strings. – Peter Cordes Sep 11 '21 at 21:53
  • 1
    meta question: should [representation] be a synonym of [data-representation] or vice versa? I didn't previously know either tag existed, but there are similar questions in both. Someone should ask on meta, I guess, unless there's an existing discussion. (I'm hoping someone who's seen either of those tags before might post a meta question so I don't have to go digging myself.) – Peter Cordes Sep 11 '21 at 21:59

0 Answers0