bitcoin-dev
[BIP Proposal] Buried Deployments
Posted on: November 17, 2016 01:41 UTC
In a discussion on the bitcoin-dev mailing list, Eric Voskuil corrected Jorge Timón's assertion that BIP30 and BIP34 were treated similarly.
Voskuil stated that duplicate transaction hashes can and will happen in Bitcoin, and that BIP34 does not prevent this. He explained that the pigeonhole principle applies to hashing algorithms, meaning that collisions are inevitable when there are more possible keys than indices in an array. Voskuil provided a link to the Wikipedia page on the pigeonhole principle for further reading. Thomas Kerin had mentioned BIP30 earlier in the thread, noting that it had received similar treatment after a reasonable amount of time had passed. However, Voskuil clarified that BIP30 is not validated after BIP34 is active because blocks complying with BIP34 will always necessarily comply with BIP30.