5 ms·

Sure, but the PoW mechanism is more fine grained, and I think it is important to point that out. Otherwise, people might indeed think that the difficulty will a

by vog 11y ago

Sure, but the PoW mechanism is more fine grained, and I think it is important to point that out. Otherwise, people might indeed think that the difficulty will always be a power of 2, which would mean that adjusting it would have been harder (and more inaccurate) than it really is.

lisper 11y ago
I dunno, that seems like a pretty small nit to me. But I suppose reasonable people could disagree.
- arcatek 11y ago
  I've worked on a coin, and this kind of inexactitude are what make it really hard to understand how the implementation works (that, and the fact that the code was hell at the beginning). If the code uses a "less-than" operator, just tell it. It makes no sense to use a different terminology (in this particular case, the "number of leading zero" concept was probably used to make it easier to understand for business guys - it doesn't work quite well).
  - vog 11y ago
    Moreover, I believe that talking about plain numbers and "less-than" actually makes explaination easier, compared to talking about strings and leading zeros. Using plain numbers, I would explain it to a layperson like this: The hash is number from 1 to 1,000,000 (actually, it's a lot larger and we are counting from zero, but you get the idea). You can't control how large or small the value is. If some people generate a hash of 10,000 or below, this means they had to try roughly 100 times, computing 100 hashes, until they found that one by chance. This is a "Proof of work". If you reduce the threshold from 10,000 to 5,000, people will have to compute on average 200 instead of 100 hashes. So the smaller the threshold, the more difficult the challenge is to solve. On top of its simplicity, this explaination has the advantage that the statistics are pretty obvious and don't even contain exponentials or logarithms, which you would have to use when talking about strings and leading zeros.
    - xxs 11y ago
      You can't control how large or small the value is. If some people generate a hash of 10,000 or below, this means they had to try roughly 100 times, computing 100 hashes, until they found that one by chance. Not to be blunt, yet after 100 attempts w/ 1% positive outcome, there is roughly 63.3% odds to actually succeed.
      - tromp 11y ago
        The expected number of trials, each with 1% of success, is 100 though.
- mikeash 11y ago
  That's probably why the comment above that pointed it out called it a "minor error."

by vog 11y ago

lisper 11y ago

I dunno, that seems like a pretty small nit to me. But I suppose reasonable people could disagree.

arcatek 11y ago
I've worked on a coin, and this kind of inexactitude are what make it really hard to understand how the implementation works (that, and the fact that the code was hell at the beginning). If the code uses a "less-than" operator, just tell it. It makes no sense to use a different terminology (in this particular case, the "number of leading zero" concept was probably used to make it easier to understand for business guys - it doesn't work quite well).
- vog 11y ago
  Moreover, I believe that talking about plain numbers and "less-than" actually makes explaination easier, compared to talking about strings and leading zeros. Using plain numbers, I would explain it to a layperson like this: The hash is number from 1 to 1,000,000 (actually, it's a lot larger and we are counting from zero, but you get the idea). You can't control how large or small the value is. If some people generate a hash of 10,000 or below, this means they had to try roughly 100 times, computing 100 hashes, until they found that one by chance. This is a "Proof of work". If you reduce the threshold from 10,000 to 5,000, people will have to compute on average 200 instead of 100 hashes. So the smaller the threshold, the more difficult the challenge is to solve. On top of its simplicity, this explaination has the advantage that the statistics are pretty obvious and don't even contain exponentials or logarithms, which you would have to use when talking about strings and leading zeros.
  - xxs 11y ago
    You can't control how large or small the value is. If some people generate a hash of 10,000 or below, this means they had to try roughly 100 times, computing 100 hashes, until they found that one by chance. Not to be blunt, yet after 100 attempts w/ 1% positive outcome, there is roughly 63.3% odds to actually succeed.
    - tromp 11y ago
      The expected number of trials, each with 1% of success, is 100 though.
mikeash 11y ago
That's probably why the comment above that pointed it out called it a "minor error."