| Glossary | Software | Open source

Open source

Open source software in the narrower sense is computer software (programs) that can be used, modified and distributed free of charge. Well-known examples of open source software are the Linux operating system [1] and the Firefox web browser [2]. Open source in the broader sense is a people-powered phenomenon that has given us unimagined opportunities for global collaboration as well as new business models.

Open source licences

In the beginning, there were the open source licences, such as the MIT License [3] or the GNU Public License v2 [4]. An open source licence regulates the rights granted to users of open source software. As already mentioned, these rights include the free use of the software, its availability in source code and the right to modify the programmes free of charge and also to pass them on. Whether a licence is a correct open source software licence is regulated by the Open Source Initiative [5], an American non-profit organisation.

The desire of creators to allow other people to freely use their work is not limited to software (and did not come about through open source software in the first place). The structure and logic of open source licences, however, has taken hold and spread to other areas. Today, for example, there are Creative Commons licences [6], which grant users of creative works comparable rights to those granted by open source software to its users, open data licences, which do the same for data, and so on.

Open source licences for the first time regulated the collaboration of programmers on software in a way that was perceived as symmetrical and fair. Anyone can use the software and these rights can no longer be denied. In principle, everyone can contribute and work for the benefit of all involved. Such collaborative work motivates many people. The motivation ranges from the fun of creative work, altruistic sharing and doing something good for others to the development of one’s own competences and the public presentation of these competences to position oneself on the labour market.

Community work

The resulting (informal) software development organisation is commonly called an open source project. The choice of word is unfortunate (but established) because an open source project is not a project in the strict sense: An open source project has no defined end date, as would be expected of a well-defined project. On the contrary, the community of contributors usually hopes for an open end, i.e. an indefinite lifetime of the software. Software, project and community usually have the same name: Thus one speaks of the Firefox software, the Firefox project and the Firefox (project) community.

Collaboration in an open source project is called open collaboration [7]. In open collaboration, everyone is allowed to participate (and no one is excluded a priori), decisions are made on the basis of the quality of arguments (and not on the basis of organisational structures), and the parties involved choose processes, methods and tools themselves (instead of having them imposed on them). In this way, open source development differs from traditional, hierarchical cooperation in companies.

Work in open source projects is almost always distributed and asynchronous (takes place at different times of the day). Developers often use the software themselves, either as part of their paid work or privately. Accordingly, it is important to ensure the quality of the work. An essential means of quality assurance is peer-based code review (principle of dual control). In important projects, no contribution is included that has not been countersigned by at least one person other than the author. Accordingly, open source developers have invented new tools for collaboration, in particular the Git software [8], which enables the implementation of simple but effective peer review.

As with licences, open source work processes are in the process of propagating into other worlds of work than software development.

Economic significance

From the point of view of companies using open source software, this software represents either parts (components) for their own products or tools (applications such as Firefox or the gcc compiler [9]) for internal work. The total cost of using open source software is not zero (that would be just the licensing costs), but it is generally much lower than that of comparable proprietary software:

Open source software reduces the risk of steadily increasing software licensing costs, which in the traditional case results from dependence on one vendor (the so-called vendor lock-in).

Likewise, open-source software prevents innovation backlogs because, in contrast to traditional software, one can help oneself to implement new functions by modifying the software.

Furthermore, open-source software reduces friction losses because it often becomes a quasi-cost-free standard to which everything fits.

Open source software has proven to be particularly effective as a means against the monopolisation of versatile software by individual manufacturers: Intel, Oracle, IBM and others, for example, have pushed the development of Linux in order to establish an alternative to Microsoft’s Windows. Without this alternative, Microsoft would be able to appropriate much of the customer’s budget, leaving little margin for further software. Investment in open source software has increased rapidly since manufacturers understood that they could use it to prevent monopolies.

What at first glance seems like a free-rider problem (everyone wants to profit by using open-source software, no one wants to invest in its development) is not: as soon as a company uses open-source software, it also becomes dependent on it: It is thus in the company’s interest to manage this dependency, for example by purchasing commercial services for the open-source software or by contributing in a controlling way.

Open Source Associations

The rapidly growing economic importance of open source software has meant that companies have had to learn to organise the management of projects in a consensual manner. In particular, disputes over trademark rights led to the establishment of so-called open source associations (“foundations”) or consortia. Examples of open source associations are the Apache Software Foundation [10] or the Eclipse Foundation [11].

In these associations, companies, but also private individuals, work together under defined governance. The use of the software is regulated by the open source licence. Any trademark rights belong to the association and are provided on a non-discriminatory basis under defined conditions. The association represents the interests of the project community, especially in court, should this become necessary.

While twenty years ago there were only a few such associations, today there are already a three-digit number, and the trend is rising.

The author of this article distinguishes here between open source associations led by software manufacturers and those led by software users. Manufacturers mostly develop components for their products collectively, while users mostly develop complete programmes for their needs. Open source software developed in this way is always non-competitively differentiated because no one can be excluded from using it. It reduces the overall costs for its users and, as mentioned earlier, can be established to prevent monopolies.

Vendor-led associations are backed by the software industry; user-led associations are backed by all industries that need software, i.e. a large part of the world economy. While software vendors have largely understood the importance of open source software, software users are still building this understanding. Examples of user-led open source associations are the GENIVI Association [12] initiated by BMW in the automotive sector, the ASWF [13] initiated by film studios for software for film processing and the Apereo Foundation [14] initiated by universities for software in education.

New business models

Open source associations have made open source projects sustainable. Start-ups have built novel business models on open source software. Essential to the new business models is that they cannot be based on the sale of open source software per se: After all, the software is available free of charge.

Customers always pay for something that is complementary to the software: for particularly efficient hardware for the software, for example for Google’s TPUs (integrated circuits) [15] for the particularly efficient application of Google’s Tensorflow (open-source software for machine learning), or for the operation of the software in the cloud, for example for Yugabyte’s cloud service [16] for the operation of the open-source software of the same name for databases, or for consulting services for the use of the software.

Non-paying users often compensate the producers of open source software in a non-monetary way, e.g. through free work, innovation or data. Many good and free bug reports help the software to mature, good ideas inspire the next generation, and data can be used profitably or sold on. For example, the data collected via the dominant web browser Chrome helps Google to run its advertising business more successfully.

The commercialisation of open source has meant that in many cases the development of open source software does not happen openly. There is no open collaboration (i.e. no open process). Instead, work is done behind closed doors and access channels (web domains, trademark rights) ensure that the competition is late and gets little or nothing of the cake.

Criticism and the future

The definition of open source software focuses purely on the licence. This has enabled manufacturers to call software that has not been openly developed but is openly licensed open source software and thus to benefit from the positive reputation of open source. This is rightly criticised by the open source world. Accordingly, people argue about expanding or changing the definition, with the aim of either diluting it (allowing novel licences) or tightening it (requiring an open process). There is no end in sight.

On the one hand, open source licences have made open collaboration possible; on the other hand, they have led to an almost unmanageable amount of work if one wants to comply with their terms. If, for example, a manufacturer wants to pass on the Linux operating system to customers, this manufacturer has to gather the relevant information for the open-source licence from many thousands of files. A special condition of some open source licences, the copyleft clause, also requires (to put it simply) that the products be made available in source code. For many years, the fear of having to unintentionally disclose intellectual property in this way scared many manufacturers away and hindered the triumph of open source.

Open source software can certainly be considered a revolution. However, two of the most historically important leaders of the revolution, Richard Stallman and Linus Torvalds, are under significant criticism for having created and led a non-inclusive, confrontational culture characterised by verbal abuse. Even today, the open source world is significantly less diverse than the already less diverse software industry. All culture change is hard, and so the situation is only slowly improving.

The increasing open source involvement of companies not from the software industry and the growing expansion of open source innovations into other fields vividly illustrate the triumphant march that open source has taken. A well-known Silicon Valley investor postulated in 2011 that software would swallow the world [17]. Open source is friendlier and lifts the world to new heights.