Jump to content

Talk:Hash function security summary

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

bits

[edit]

Suggestions: tables on this page would greatly benefit from having a column expressing the number of bits in the hash function outputs. This permits comparison of the ratio "security/space used". In almost all cases, when the algorithm is sound, using more space will create more security, which is connected directly to a main point of hashing, saving bit space in a secure description. 192.235.78.16 (talk) 05:14, 22 November 2016 (UTC)[reply]

Merge

[edit]

I'm curious if there is some way to either split a few pages or some method to keep pages with the same content up to date. For instance there is at least: Cryptographic_hash_function Comparison_of_cryptographic_hash_functions Hash_function_security_summary

They aren't always all in sync with each other and that's not to mention the pages for each hash function. Thoughts? Quelrod (talk) 18:24, 31 July 2010 (UTC)[reply]

Merging this page with Comparison of cryptographic hash functions would be a first small step. I don't see a reason for having two comparison pages. Updating the comparison pages more conservatively might help too. I.e., has it been verified that the 251 against SHA-1 really has this complexity? The author clearly states in the paper that the result is based on heuristics. On the SHA-1 page it is always possible to include papers with uncertain results with a short comment. On a page with summaries, however, this leads to oversimplification. 85.1.63.54 (talk) 06:10, 1 August 2010 (UTC)[reply]
@Quelrod: I have merged all content from Comparison of cryptographic hash functions table to here. I think this format allows a more detailed approach than the other comparisons. It prioritizes the number of broken rounds over time complexity — an attack that doesn't break all rounds isn't actually a "successful" attack. And it allows memory complexity to be taken into account too.
I used personal judgement in some cases. For example the Tiger collision attack [1] from there actually only speaks about pseudo-collisions, which I think isn't the same as a proper collision. The novel MD4 attacks reported in [2] are interesting, but if you take into account their memory or precomputation requirements, then the original 2008 attack appears cheaper. I would be happy to discuss these if someone has an opinion.
As for IP user's comment about verification, such confirmations don't usually come about. I think the best we can do is to make sure that the papers come from reputable journals or conferences, which have peer review processes in place to prevent errors.
I also proposed a merge for the Cryptographic hash function#Cryptographic hash algorithms table; I think the attack columns should be merged here too. It makes the same mistakes as the other comparison, and it even colors unsuccessful attacks pink. The rest of the table should probably go to Comparison of cryptographic hash functions. -- intgr [talk] 20:12, 21 October 2014 (UTC)[reply]
I have now completed both merges, no more duplication. :) -- intgr [talk] 19:27, 23 October 2014 (UTC)[reply]