4 ms·

This doesn't seem like a joke, it seems like a guy outright spreading misinformation. Not sure what was the point of the UUID preamble, his "joke" has nothing t

by mdoms 5y ago

This doesn't seem like a joke, it seems like a guy outright spreading misinformation. Not sure what was the point of the UUID preamble, his "joke" has nothing to say about UUIDs.

rootusrootus 5y ago
Only if people take seriously the posts random anonymous strangers make on r/programmerhumor. The real source of disinformation here is whomever decided to repeat the dry humor without the context of it coming from a joke subreddit.
- pessimizer 5y ago
  Some people will believe everything they read, and they are friends with people who will believe everything they heard from a friend.
TheDudeMan 5y ago
And his link to the Google collision was not an accidental collision, like the rest of his ramblings were discussing.
- jameshart 5y ago
  Right - that’s a SHA1 collision, which has nothing to do with the possibility of a UUID collision. I worry that some people might take away from this that UUID collisions are actually something to worry about (nonsensical kindergarten cubby analogies and so on). Really, truly, UUIDs can be assumed unique. We’re not relying on luck.
  - zamadatix 5y ago
    UUIDs can be assumed unique because we are relying on luck, not the other way around. The cubby example is actually a pretty good example of why the luck is on your side to the point you should never worry about it. The SHA1 collision not as much.
  - hanselot 5y ago
pbhjpbhj 5y ago
The humorous musings are, despite the unlikelihood of collisions the author was pondering them for UUID (see backpack cubby analogy), this situation came up and then pondered what might have happened with a similarly unlikely collision -- byte identifiers between Roku and Pokemon. It's silly, it's in a humour forum.
- vlovich123 5y ago
  The probability of getting a UUID collision is different from an intentional attempt to find a SHA-1 collision. One is guessing a truly random ~128 bit number (really - it’s effectively the output of a CSPRNG). The other is exploiting a weakness in the construction of a 160-bit value that is intended to be a secure mapping of input -> output. It’s not even remotely the same problem and saying evidence of a sha-1 collision indicates anything about a UUID collision is plain wrong (and that part isn’t on a humor forum).
  - pbhjpbhj 5y ago
    >saying evidence of a sha-1 collision indicates anything about a UUID collision is plain wrong // No one said that, so ...
    - vlovich123 5y ago
      Did you read or skim? > The odds of two UUIDs being generated with the same values on the same system is very low …snip analogy description > Except it could still happen. You usually get some hand-waiving from software engineers when you say this, but... collisions could still happen. Google found one. That last bit is a link to Google brute forcing a sha1 digest

by mdoms 5y ago

This doesn't seem like a joke, it seems like a guy outright spreading misinformation. Not sure what was the point of the UUID preamble, his "joke" has nothing to say about UUIDs.

rootusrootus 5y ago

Only if people take seriously the posts random anonymous strangers make on r/programmerhumor. The real source of disinformation here is whomever decided to repeat the dry humor without the context of it coming from a joke subreddit.

pessimizer 5y ago
Some people will believe everything they read, and they are friends with people who will believe everything they heard from a friend.

TheDudeMan 5y ago

And his link to the Google collision was not an accidental collision, like the rest of his ramblings were discussing.

jameshart 5y ago
Right - that’s a SHA1 collision, which has nothing to do with the possibility of a UUID collision. I worry that some people might take away from this that UUID collisions are actually something to worry about (nonsensical kindergarten cubby analogies and so on). Really, truly, UUIDs can be assumed unique. We’re not relying on luck.
- zamadatix 5y ago
  UUIDs can be assumed unique because we are relying on luck, not the other way around. The cubby example is actually a pretty good example of why the luck is on your side to the point you should never worry about it. The SHA1 collision not as much.
- hanselot 5y ago

pbhjpbhj 5y ago

The humorous musings are, despite the unlikelihood of collisions the author was pondering them for UUID (see backpack cubby analogy), this situation came up and then pondered what might have happened with a similarly unlikely collision -- byte identifiers between Roku and Pokemon. It's silly, it's in a humour forum.

vlovich123 5y ago
The probability of getting a UUID collision is different from an intentional attempt to find a SHA-1 collision. One is guessing a truly random ~128 bit number (really - it’s effectively the output of a CSPRNG). The other is exploiting a weakness in the construction of a 160-bit value that is intended to be a secure mapping of input -> output. It’s not even remotely the same problem and saying evidence of a sha-1 collision indicates anything about a UUID collision is plain wrong (and that part isn’t on a humor forum).
- pbhjpbhj 5y ago
  >saying evidence of a sha-1 collision indicates anything about a UUID collision is plain wrong // No one said that, so ...
  - vlovich123 5y ago
    Did you read or skim? > The odds of two UUIDs being generated with the same values on the same system is very low …snip analogy description > Except it could still happen. You usually get some hand-waiving from software engineers when you say this, but... collisions could still happen. Google found one. That last bit is a link to Google brute forcing a sha1 digest