EC FLOSS Report Installment #3: Where are the FLOSS Developers?

by on February 20, 2007

Based on my prior post that analyzes Section 6 of the European Commission’s FLOSS report, we know that there is a growing and in many cases strong private and public sector demand for certain FLOSS applications. This leads the authors to the question of supply? Where are the FLOSS developers?

Based on two developer surveys (MERIT/FP5 and FLOSS-US
Stanford), the report argues that more than three fifths of the worldwide
FLOSS developer community live in the EU.
One fifth live in North America and another one fifth live in other countries (p.37). The report relies on this and other data to persuade the EC to adopt an industrial open source policy.

While the authors may be overstating the point for effect
(the different data sets do not always show Europe leading and many Asian
developers are not counted), it is pretty clear that Europe is a leader in global community of FLOSS developers. 

However, as I say below the fold, the authors gloss over some important developer community distinctions that may harm their later conclusions: 

  • Committer vs. Submitter–The authors may actually understate the number of globally active FLOSS developers by limiting their definition of an “active core developer” to only one that commits software (Figure 20, p.39). Generally speaking, committers are developers that are given write access to the code repository. There may be many active developers that submit code, but do not possess the access privileges of committers. 
  • Committer vs. Real Contributor – The report weighs each committer’s contribution the same. Yet, some projects are important while others are trivial, and some developers commit thousands of lines of code while others commit only a few. Casual developers who have only contributed a few lines of code do not provide the same value or opportunity to the European economy that a Linus Torvalds does. Without considering these distinctions, it is difficult to analyze the EU’s FLOSS production capabilities. 
  • Contributor vs. Leader – Most FLOSS is written by a select group of developers (and often employed to do so). In the Summary Findings for Section 6, the authors state that “almost two thirds of FLOSS software is still written by individuals.” This is misleading at best. Individuals participate on a wide basis, but most code comes from only a small, select group of programmers, many of whom are paid by their company to write code. As Andrew Morton is on record of saying:

We take the example of the Linux
kernel project. Over the past three years, there have been 38,000 changes to
that tree–38,000 pages. One thousand contributors, one thousand individuals have
made those 38,000 changes. But 20 contributors have made half of those changes.
So the remaining half of 38,000 lines came from the other 980 people. So it’s a
very non-uniform distribution of contribution rights there.

Furthermore, JBOSS CEO Marc Fleurie
has reportedly said that 80% of the JBOSS kernel was written by 20 developers,
all employed by JBOSS. 

From the Silly Inaccuracy Department: Linux is not a
database!
–The authors describe the growth of the open source database
market by stating that “Linux had the greatest momentum” but that Oracle
remained at the top position (p.34). 

In conclusion, this section highlights undeniable trends in
the growth of free and open source software. However, the report’s data can and
should be improved upon as the EC continues its ongoing consideration of a new Airbus-like industrial policy based on FLOSS for the European ICT market. 

Next, I’ll write about Section 7, which analyzes the direct
economic impact of FLOSS on ICT sector employment. 

Comments on this entry are closed.

Previous post:

Next post: