As wildlife scientists and professionals, most of us rely oncomputer software in our daily work. We invest a large pro-portion of our time in learning how to use different kinds ofsoftware effectively to do our jobs. Choosing which soft-ware to learn in the first place is therefore an importantlong-term decision. Many of us rely on commercial soft-ware such as Word, Excel, Outlook and SPSS or SAS whenwe analyse data, correspond with our colleagues by e-mail,and write papers.In this paper, we will briefly discuss some importantaspects of commercial software in general. More impor-tantly, we will introduce the reader to a specific alternativeto commercial software known as free software. Our im-pression is that the awareness of this phenomenon is muchsmaller among biologists than in other fields of science suchas mathematics, statistics, physics, and engineering. Freesoftware is software which is distributed freely, typicallyover the Internet, under a licence known as the GNU Gen-eral Public Licence (GPL) (Free Software Foundation, Inc.,1991). Like commercial software licenses, this licence re-lies on standard copyright law. The important difference isthat the recipient of the software is granted a number of ad-ditional rights by the copyright holder which permits the re-cipient of the software to make changes to the source code,provided that any modified version of the source code isalso made freely available under the the same license. Thiscontract therefore guarantees that any improvement made tothe software becomes freely available to the software’s ini-tial author. This potential reward becomes one of the mainmotivations for releasing the software as free software inthe first place.This licencing regime has dramatic consequences on theeffectiveness of the software development process. To un-derstand why, we need to understand approximately howthe process of complex software development typicallyworks in general. When an architect designs a new building,almost all the details of the design are typically completedbefore any building work is initiated. A common miscon-ception about software development is that software is de-signed in a way similar to this. In reality, however, mostsoftware, both commercial and free changes through grad-ual small incremental modifications of the source code afterreaching a certain level of complexity. Complex software is
[1]
Thomas D. Sandry,et al.
Introductory Statistics With R
,
2003,
Technometrics.
[2]
Russell B. Millar,et al.
BUGS in Bayesian stock assessments
,
1999
.
[3]
Leslie Lamport,et al.
Latex : A Document Preparation System
,
1985
.
[4]
James N. M. Smith,et al.
Bayesian meta‐analysis of demographic parameters in three small, temperate passerines
,
2000
.
[5]
N. Stenseth,et al.
Cycles and trends in cod populations.
,
1999,
Proceedings of the National Academy of Sciences of the United States of America.
[6]
Thomas Petzoldt,et al.
R as a Simulation Platform in Ecological Modelling
,
2005
.
[7]
D. Spiegelhalter,et al.
Bayesian Analysis of Realistically Complex Models
,
1996
.
[8]
Peter Green,et al.
Markov chain Monte Carlo in Practice
,
1996
.
[9]
Tobias Oetiker,et al.
The Not So Short Introduction to L A T E X2
,
1999
.