Open source fans are not inclined to share. Hitchhiker's Logic

Original author: Matt Asay
  • Transfer
This article is a translation of Matt Esey's article, “Open source's ardent admirers take but don't give. Free riders' self-defeating logic, ”published August 13, 2010 on The Register .

Matt asayOpen source software has long been something commonplace. Therefore, it is all the more pity that many of his most ardent admirers still do not understand the principles underlying it.

Ask, for example, the CIO team how Accenture recently did : why are they so zealously deploying open source software - and they are really zealously deploying it because:
  • 50% of respondents are “fully committed” to open source software;
  • 69% plan to increase investment in it;
  • 38% generally plan to transfer all critical software for their business to open alternatives in 2010.

And what do you hear in response?

Open source software is characterized by better quality (76%), higher reliability (71%) and lower support costs (71%). These numbers are also confirmed by Forrester polls .

Great, right? Perhaps. Until we add another spoon with numbers to this barrel: only 29% plan to return their changes in the code to the relevant communities.

Generally speaking, this does not mean that everything is bad. In the end, the ideology of open source software does not oblige everyone who uses it to return the contribution to the community, as long as there is a critical mass of developers interested in developing projects that are sufficient to ensure that these projects are not bent.

But the question arises: what do these CIOs expect from open source software?

Imagine a situation: an enterprise makes changes to open source software, but does not publish changes, and these changes do not fall into the main branch. This means that from now on it must support this software on its own. And in this case, not only increases the cost of maintaining the software, but also disappear, or at least reduce the benefits of quality and reliability of the code.

Admittedly, not all developers are equal, and you cannot guarantee that your Linux kernel fixers are as capable as Google’s Ted Ts'o or James Bottomley of Novell.

It is curious that these two were arguing about this last week at LinuxCon in Boston, as The Register has already reported . Google displeased: it allegedly forked the Linux kernel for its Android; the fact that he did not return the changes made for Android back to Linux and thereby separated from him and from the vast Linux community; in that it reduces overall performance and degrades compatibility between the two branches.

Although Google promises to honestly return everything Android has gained, it doesn’t seem to be in a hurry. Why's that? If Google is not heavily burdened with additional costs - and he, obviously, is not burdened with them, he callsAndroid development costs are "insignificant" - then he probably would only benefit from the publication of the Android code ... and yet he does not .

As for Google, he, of course, did not hesitate to provide Accenture with information about the money saved on maintaining the software. He almost certainly believes that he is able to provide equal or even better reliability and quality than the Linux kernel community. Google has ruled the kernel for many years for its search engine and other products and probably knows it thoroughly.

In other words, Google is not cunning when it comes to using open source. They understand the costs incurred by their own development of Linux, and are ready to bear them.

Still, I'm curious if the CIOs surveyed by Accenture are aware of the conflicting views on open source software: what they expect to get from it and what they are going to pay for it.

In general, is this a problem? Not.

Linux Foundation Executive Director Jim Zemlin told me at a LinuxCon lunch:

It always ends with the code being returned to the community - even when a project starts with a fork of an existing project and is developed alone. Supporting a project on its own over time becomes simply irrational, and companies eventually realize this.


The sensible CIO must recognize that open source software always wins when a new member joins the game. According to Zemlin, they do not always immediately understand this, but in the end they still give up, which is confirmed by examples from IBM, Intel and others.

Of course, there is a certain benefit from a simple “hitchhiking” at the expense of the community. But the benefit increases many times if the enterprise not only exploits the community, but also actively participates in it.

Matt Asay is Canonical's chief operating officer for commercial operations at Ubuntu. Engaged in open source software for more than a decade, Esey served as general manager for the Americas region at Alfreso, vice president of business development, and also helped launch the open source business at Novell. He is an honorary member of the board at the Open Source Initiative (OSI). His column entitled “Open ... and Shut” goes to The Register every Friday.

PS Good health to you, Oracle! :)

Also popular now: