Talk:Jenkins hash function

signed vs unsigned

It should be noted that for the += key[i] step it is a difference wether key is a char* or unsigned char*, and in my experience the unsigned version has a slightly better avalanche behaviour.

Who is Bob Jenkins?

I'm glad he decided to contribute so much, but honestly it seems like he is just plugging his website. Unless his work has been published in a peer-reviewed journal, I don't think it should be continuously cited throughout this article. A simple mention of his site in the external links section would suffice. --Kibblesnbits 17:16, 13 March 2007 (UTC)

Bob has produced some very good research articles in hash functions. We who do hash function research are often in the industry, and do not have easy access to academic publications. MegaHasher 06:31, 18 August 2007 (UTC)

I second that, I find code using the jenkins hash sometimes in other applications and libraries, and also jenkins seem to have made the research on avalanche behaviour quite popular. Also, if you surf aronud on his (horribly organized) website you will find some publications

I agree that the article is pretty weak, however I see it being used in HBase (HDFS/Hadoop) and people have ported versions to python, java, and ruby. I think it's important enough to keep, BUT the article needs clarification and expansion. Cgthayer (talk) 16:17, 12 May 2010 (UTC)

Perl has used Jenkins one-at-a-time hash function for years. IMO for that reason alone it deserves a reference. — Preceding unsigned comment added by 212.64.88.14 (talk) 23:03, 7 December 2012 (UTC)

The Embarcadero Delphi runtime library contains an implementation of it nowadays as well.--87.180.121.81 (talk) 19:09, 9 June 2016 (UTC)

License ?

This page reproduce the algorithm for the "One-at-a-Time" hash, is it allowed to ?

What is the license applicable to the Bob Jenkins hash algorithms ? Ydroneaud (talk) 08:42, 3 April 2013 (UTC)

The very first link in the references section[1] says "The code given here are all public domain.". Is this something we need to emphasize more in this article? --DavidCary (talk) 23:43, 29 May 2013 (UTC)

Notability

Intgr, regarding the notability and sourcing of this article:

• I added a reference to a conference paper (44 cites on GS) that discusses one of these hash functions in some detail.
• Benchmarks and short comments on SpookyHash occur in the SipHash paper.
• The old Jenkins hash was used to solve Kalah. This paper discusses some details of the hash function and argues why, for this game, it outperforms Zobrist hashing.
• Netfilter uses Jenkins hash. It was implemented to mitigate the risk of hash flooding.

I'll put some more of this into the article. Please reconsider the notability template. QVVERTYVS (hm?) 20:39, 11 June 2015 (UTC)

@Qwertyus: The SPIN paper is good because it actually discusses one of the variations in length, which can be used to expand the article with relevant content. But the other two are just name-dropping the function and don't qualify as "significant coverage" by WP:GNG. The notability requirement is there in order to make sure that sufficient reliable sources exist, to base the article on (WP:WHYN).

Building a list of software that uses this hash isn't an improvement to the article IMO; it's not much more useful than a "list of products built using M5 screws". I think you should concentrate on improving content about the Jenkins hash itself and find sources that help you do that. -- intgr [talk] 14:49, 15 June 2015 (UTC)

@Intgr: I actually picked exactly those sources that commented on their choice of hash function. I think notability is now well-enough established, and I'm removing the templates. QVVERTYVS (hm?) 15:06, 15 June 2015 (UTC)

@Qwertyus: Notability is not about puffery. The notability guideline, WP:GNG, requires sources with "significant coverage" and two of your three sources fail to provide that. And you didn't even begin to solve the primary sources issue: the majority of the article still relies on primary sources. -- intgr [talk] 15:14, 15 June 2015 (UTC)

Feel free to restore the tags, then; you're right about the other hash functions, but I don't consider what I posted to be puffery. Let's get a third opinion. QVVERTYVS (hm?) 20:33, 15 June 2015 (UTC)

I removed this from the third opinion noticeboard because notability doesn't appear to be an issue anymore. If any other disagreement(s) come up between you two, feel free to relist. Erpert ^{blah, blah, blah...} 01:27, 4 July 2015 (UTC)