Talk:Linear feedback shift register
From Wikipedia, the free encyclopedia
Hi, I think the paragraph submitted by 69.238.5.10 on 19:08, 22 June 2006 is not meant to be part of the content for this article. Instead, the submitter is referring to an error in the animation image.
"The animation below is erroneous. In one stage of the cycle the values of the taps do not correspond to the values stored in bit positions 14 and 16. The values of the taps rather than the values in the bit positions are then used to calculate the code, thus propagating the error into the feedback. This occurs when the last 3 bits are 100."
Matt Crypto, can you update the animation and republish? -daniel.kho
- Oops, silly mistake, sorry about that. I can't access the sources images right now, but I'll try and get to them shortly. — Matt Crypto 13:22, 14 September 2006 (UTC)
Maybe Rob.derosa is correct. The runlength (or pattern length) of a 6-bit LFSR should be 26 − 1 = 63 rather than 26 − 1 = 32.
"In one period of a maximal LFSR, 2n − 1 runs occurs (for example, a six bit LFSR will have 63 runs)."
I suggest introducing hyphenation for clarity:
"In one period of a maximal LFSR, 2n − 1 runs occurs (for example, a six-bit LFSR will have 63 runs). Exactly 1 / 2 of these runs will be one-bit long, 1 / 4 will be two-bit long, up to a single run of zeroes (n − 1)-bit long, and a single run of ones n-bit long." -daniel.kho
- No. In one period of a maximal LFSR (which is 2^n - 1) it produces 2^n-1 BITS. You cannot have 2^n-1 RUNS in 2^n-1 BITS (well, unless you have only 1 bit runs: "010101010101010101...", which doesn't happen). I suppose you will have 2^(n-1) runs? —The preceding unsigned comment was added by 195.212.29.163 (talk • contribs).
[edit] Where is the period of maximal LFSR stated?
Hi, I don't see where it is stated that the LFSR maximal period is 2^N,with N the number of state bits, am I missing something here?
[edit] Galois source code
I recently cleaned up the C source code for the Galois LFSR. However, I am considering removing everything but the actual implementation of the LFSR (2-3 lines of code), because I think most people would find that more useful (and less confusing). I am also unsure how "optimized" the code should be. Any opinions? -Ufretin 14:40, 16 January 2007 (UTC)
[edit] The tap values in a maximal LFSR will be relatively prime
Is there a proof for this statement? How shall the tap values be taken? The polynomial 0133 is a maximum length polynomial, if it is written as x^6 + x^4 + x^3 + x^1 + 1, the tap values would contain 6 and 3, which have 3 as a common factor. 84.161.181.208 19:48, 28 January 2007 (UTC)
