What is Digg?Sponsored by Dragon Age: Origins
Can't get enough Dragon Age: Origins? Play the flash game. view!
DragonAgeJourneys.com - Play the free companion flash game to Dragon Age: Origins.
27 Comments
- Cupantae, on 12/01/2008, -2/+17Fascinating. I have been reading about the nature of open source development recently, and I would offer some suggestions. The reasons I believe the popularity of packages would follow Zipf's Law are:
1. There is no minumum quality for any given package, so there exists many unpopular ones.
2. A very small proportion of the projects have commercial backing, e.g. from Mozilla and Sun, and so, a small number receive a disproportionate amount of attention.
3. Sense of community is very important to open source development, and so a community attracts further improvement. This is, in a sense, a circular argument, but what it leads to is strong backing of a few breakaway projects.
4. The packages are freely distributed, so people are free to try out all the alternatives for any given application, so people tend to settle on one which is considered the best.
As an added bonus comment, I don't quite agree that open source is "based on altruist contributions by programmers”. This does not tell the full story at all. The reasons for open source development include desire for the end product, political views, improvement of professional standing, sense of belonging and worth in a community and, let's not forget, almost half of the work is done by paid programmers. - Rapax, on 12/02/2008, -0/+9This got me thinking about how using, or deliberately avoiding, Zipf may be a useful tool. If most literary works follow Zipfs law, then it stands to reason that we're used to reading text that follows Zipf. So if you write something that deliberately breaks that law, for instance by doing a Zipf analysis, then replacing common words with synonyms just enough to flatten out the distribution, it would probably register in the subconscious of the reader as being 'slightly off'. Maybe this could be a useful tactic when you want your text to stand out from a bunch of essentially identical texts, e. g. in a job application.
- nickpick, on 12/02/2008, -2/+6...again.
- HonoredMule, on 12/02/2008, -0/+4Odd that there's no similar law regarding the pedants. At any rate, I'm here.
Strictly speaking, we didn't break Godwin's Law, we satisfied it. - asgardshill, on 12/02/2008, -1/+3This means Zipf to me - buried.
- orlyfactor, on 12/02/2008, -1/+3Do this one...for the Zipper!
- KingCritter, on 12/02/2008, -0/+2http://en.wikipedia.org/wiki/Technological_singula ...
The future looks anything but boring. - strangeman, on 12/03/2008, -0/+2Most people apparently didn't understand the article (especially the commenters on physorg). This is not about Linux but about the "why and how" of Zipf's law. They just used data from Debian to study the law.
- smotpoker, on 12/02/2008, -0/+2Also, don't forget that the mainline Linux kernel is not more than 1/4 of all FOSS development (I would guess). So probably 1/8-1/4 of packages (including small distro-specific patches) of any Linux-based FOSS system has corporate sponsorship.
- smotpoker, on 12/02/2008, -0/+2"I don't quite agree that open source is "based on altruist contributions by programmers”"
It's called reciprocal altruism which exists any many socially advanced species. Some people really do feel good when they help others and do so primarily for that purpose. You are right much occurs because of personal benefit but the desire to help others does have a strong influence.
"let's not forget, almost half of the work is done by paid programmers"
This is misleading, I believe. Half of what work and who pays them? Most corporate backing on the Linux kernel has only been added the last couple of years and focused primarily on virtualization fronts. There are plenty of other paid programmers who FOSS apps but most do not get paid to work on them. They make an app (or plugin/patch) that they need, decide others could use it and decide to share it. Sometimes they may be paid by whoever they work for while doing it but ultimately they aren't being paid to contribute to FOSS at all in most cases. - inactive, on 12/02/2008, -0/+2Very interesting subject, msaleem, but I don't understand what "mechanism" influences humans to use words at frequencies related to zipf's law while at the same time governing the population distributions of large cities and interconnectivity of debian packages. Judging by the article, it doesn't appear the author knows either. Here is the actual paper -> http://arxiv.org/PS_cache/arxiv/pdf/0808/0808.1828 ...
- Colindean, on 12/02/2008, -1/+2Fourth comment in and we've already broken Godwin's Law. Good job, folks.
- nickpick, on 12/02/2008, -0/+1Good to have an interesting article in the Linux section for a change.
- CharlesW1970, on 10/30/2009, -0/+1so in short, randomness is not random?
- m0llusk, on 12/02/2008, -0/+1The last phrase from the first sentence of the paper:
" ... driven by the mechanism of proportional growth. ... "
Proportional growth is highlighted throughout the paper, and the Executive Summary that follows the introduction starts with prominent references to Gibrat's Law of Proportionate Growth.
There is a great deal of room for more data and analysis, but the paper appears to assert that proportional growth is the core mechanism involved and that further work with proportional growth may reveal precisely how it consistently results in inverse power rule distributions. In plainer language, it isn't that the authors don't know either, but that they are not satisfied with available characterizations of the core phenomena involved even though they make use of what Gabaix and Gibrat have done. Without this result it is not clear that proportional growth should or even could be extended to explain Zipf distributions. - orlyfactor, on 12/02/2008, -1/+2Come on no Airplane! fans out there? His name was George Zipf too - bah!
- Frayed_Knot, on 12/05/2008, -0/+1What do you mean "we"? ethana2 did it.
- PennFarmer, on 12/02/2008, -0/+1Stone Age Open Source FTW
- PennFarmer, on 12/02/2008, -0/+1I would have to wonder though what the effect would be. Would it stand out enough to stick in their mind, or would it stand out in such a way as to be annoying? Another possibility is that they might feel that the author couldn't write.
On the other hand, it might lead to the opposite feelings. The reader might feel that the writing was fresh and original and the author had a good grasp of the language. I would be curious to see a study done on this. - yahoofrom, on 12/03/2008, -0/+1Zipf's law is like normal distribution! it's everywhere!
- ethana2, on 12/02/2008, -5/+5You know who else follows Zipf's law? Nazi Germans.
- sgtcaboose, on 12/02/2008, -2/+2Ive seen a tonne of these graphs recently showing that everything we do will eventually stop (i.e. medical milestones, computer speed, growth of population etc. etc.)
The future looks boring. - djfang, on 12/02/2008, -1/+1This is intriguing. This is exciting as open source software allows for collaboration and this will only increase exponentially as we become more interconnected. Great achievements will be made across the spectrum but since people are working together, this may be bad too. There are those tht may be discouraged due to the lack of recognition/incentive.
- obchunky1, on 12/02/2008, -1/+0It is an interesting principle that I'd never heard of. Why should things (cities, economies, papers, Linux etc) follow this sort of pattern. I can see how in seperate instances there would be these patterns, but for so many different events to follow the same pattern is hard to comprehend.
- thecheatah, on 12/02/2008, -3/+1This explanation doesn't explain it at all. There are so many other models which would fit your explanation better. Something along the lines of a heavy tailed distribution.
OOh.. and on top of that it has nothing to do with the quality of the packages, its the number of dependencies within a package.
I guess people on digg are really dumb. (I am not talking about the author of the comment, but generally the people who over looked this.) - oilcan, on 12/02/2008, -5/+0probability also predicts that most of humanity willl die soon because we are so far above carrying capacity collapse is inevitable. celebrate linux while you can, the new stone age is a'comin bitches!
- mctom987, on 12/02/2008, -10/+12009—Year of the Linux Desktop
© Digg Inc. 2009
— Content created and posted by Digg users is