2
$\begingroup$

I was wondering if the encoding process in coding theory is a hash function?

Are hash function and what is studied in coding theory basically the same thing?

Thanks and regards!

1 Answers 1

1

No. Coding theory deals with encoding and decoding of information, which usually implies that all the information is retained. (There's lossy coding, but still the intention is to retain as much of the information as possible within a given amount of space.) By contrast, the idea of a hash function is to distill a possibly large amount of information into a small amount of information that's just large enough for it to be unlikely that two hashes of different objects coincide.

  • 0
    Thanks! I still don't understand well. As far as I know, both encoding and hash function try to map different pieces of information to different integers, with possibly other objectives.2011-07-11
  • 0
    @Tim: Hashing only tries to avoid having "too many" collisions.2011-07-11
  • 1
    @Tim: This is true, but in encoding, the output should be different for *all* possible inputs, whereas in hashing it only needs to be *probably* different for the *actual* inputs. You should be able to *reconstruct* the original information from the encoded information, whereas that's not possible from a hash.2011-07-11
  • 0
    @Tim: Not only should you be able to reconstruct the original information from an encoded version. The process should be very easy for encoding to make any sense (as in effortless ten million bits per second to produce your digital HDTV video by a relatively simple device). OTOH when hashing (assuming that you mean hashing aimed at cryptographical security) reconstructing should be computationally infeasible. Furthermore, encoding always increases the length of the message, hashing (often but not always) decreases it.2011-07-11
  • 0
    Encoding doesn't always increase the length of the message -- many types of encoding -- Huffman encoding, arithmetic encoding, ... -- are designed to decrease the length of a typical message (of course at the expense of increasing the lengths of atypical messages).2011-07-11
  • 0
    @joriki: Ok. I would call that source coding/compression, but I guess you're right. My exposure is so heavily on the channel coding/FEC side that I tend to forget those. I've mostly seen Huffman *coding* as opposed to *en*coding, but both appear to be in use.2011-07-11