Tactical Technology Collective logo

The Politics of Open Source Adoption, NGOs in the Developing World

SSRC Report by Gabriella Coleman, University of Chicago

Introduction

NGO and non-profit sector interest in FOSS began to emerge roughly three years ago, in step with the maturation of a number of prominent FOSS solutions, the growth of private sector and government-sponsored adoption, and the general—and widely publicized—perception that FOSS constituted a viable, non-commercial alternative to Microsoft domination of the software market. Online literature, journalistic coverage and (limited) scholarship on the topic reveals palpable excitement about FOSS's potential—often in ways that express and pull together both pragmatic and political motivations./1/ A number of prominent NGO-based FOSS success stories (both inside and outside the U.S.) have played a large role in widening FOSS enthusiasm in the sector./2/

This enthusiasm belied—and sometimes ran aground against—the considerable difficulties many organizations faced (and continue to face) in transitioning to FOSS. Greenpeace, one of the largest and most fiscally sound non-profits in America, was unable to meet its posited goal of migrating their desktop machines to FOSS by the end of 2003. More typical NGOs, with limited budgets and small or non-existent IT personnel, face greater hurdles, particularly with respect to Intranet and desktop software./3 These challenges are usually magnified in the developing world, where institutional resources and society-wide IT infrastructures are inconsistently available or scarce. Adoption is further are hampered in countries (including nearly all developing countries) where the private sector in open source technologies is underdeveloped. Such a sector is indispensable to providing the local tech support and competition in services that can put FOSS solutions on level terrain with aggressively expanding commercial players, namely Microsoft.

In the past three years, these barriers have diminished in many countries, and show signs of continuing to do so as the FOSS developer community expands, and as it becomes more responsive to the needs of non-technical users. This progress has underwritten continued and expanding NGO interest in FOSS technologies, including in the poorest and most challenging locales. This chapter is an exploration of the unique sectoral conditions underlying FOSS adoption among NGOs, focusing on an account of emerging intermediaries (often NGOs themselves) who promote FOSS and facilitate its adoption in the NGO sector.

To date, the developer community has played a limited and mostly ad hoc role in promoting FOSS in the NGO sector. Although there is a powerful and high-profile form of FOSS promotion associated with a number of FOSS founding figures (among them, Bruce Perens), there are only a few organized, lower-level, developer-driven organizations, such as GeekCorps and Free Geek, that bridge the gap between FOSS principles and FOSS implementation in the non-commercial, non-educational sector, where FOSS business opportunities are limited. Instead, this advocacy work has been taken up in increasingly organized fashion by a handful of American and European NGOs. Organizations such as Tactical Technologies Collective (TTC), my main example here, have developed a strategy of building and connecting grassroots FOSS resources and expertise within the NGO sector as a basis for extending FOSS adoption.

These organizations, networks, and strategies are recent, modest in scale, and still very much in formation. FOSS advocacy in the NGO sector is still primarily characterized by “evangelism” rather than by active adoption. Much of the activity of TTC and similar NGOs is oriented toward the nuts and bolts issues of explaining the general viability and advantages of FOSS, especially around issues of security, customization, and localization. Within this context, however, TTC works to place FOSS within an important and, to date, very resonant conceptual framework, emphasizing the symmetry between the professional ethos of many NGO institutions and the underlying philosophy of openness and collaboration that drives FOSS development./4/

Tactical Technologies and the Creation of a Fledging Network

An Amsterdam-based NGO founded in 2001, Tactical Technologies Collective’s mission is to improve the IT capabilities of NGOs working in the developing world./5/ Although TTC promotes proprietary technologies when necessary or appropriate, FOSS is the backbone of their efforts.

TTC focused its initial efforts on building a network through which NGOs could learn about FOSS—something almost entirely absent from the NGO sector before 2001. This has involved building an infrastructure of people, knowledge, resources, documentation, intermediaries, vocabularies, technology, and institutions through which FOSS can become a more visible and attractive option for NGOs.

Source Camps: Challenging the “Conventional Wisdom” on FOSS

Instead of building a “virtual” network of partner institutions, TTC chose to ground its network in a series of “source camps”—week-long venues for intense face-to-face interaction among the relevant stakeholders (NGO professionals, FOSS developers, technology activists seeking to promote FOSS, technology consultants, local IT professionals). These actors were selected for their capacity to strengthen what TTC calls FOSS's “local practical implementation capacity.” The first of the camps, “Summer Source,” was held the last week of August 2003 on the island of Vis off of Croatia's Adriatic coast. It focused primarily on Eastern European NGOs.

Because most source camp NGO participants have little or no direct experience with FOSS, the camps are oriented toward baseline education about the differences between FOSS and proprietary software, including social, legal, and technical differences. Some of the initial work, however, involves addressing widely-circulating nuggets of FOSS “conventional wisdom,” especially those that exaggerate the difficulties or the benefits of FOSS adoption. The question of ease of installation is a frequent example: some see FOSS technologies as excessively difficult to install; others have misconceptions about its simplicity. In practice, the difficulty of installation depends on a number of factors: the particular software application (some applications, like Mozilla, are far easier to install than others, such as task tracking software); the depth and quality of documentation (FOSS documentation has improved dramatically in the last few years, although the amount and quality of documentation still varies tremendously from project to project); and various externalities such as the cost and speed of Internet connections and the availability of local tech-support.

A second frequent misconception relates to the total cost of software ownership, or TCO. Because downloading and installation of FOSS software is often free, many conclude that it has a nominal TCO. While licensing fees are small or non-existent, the total cost of ownership for FOSS can be higher than for comparable proprietary software. In developed countries, a significant share of the FOSS/Microsoft debate is devoted to competing TCO studies, especially the long term costs of licensing. In developing countries, TCOs can vary for reasons that have little to do with these monetary costs or with the capabilities of the software itself. The existence of a thriving black market in pirated (proprietary) software, for example, often supports local IT expertise trained in that software—typically in industry standards such as Microsoft software. Some NGOs keep their operating costs down by using pirated software and drawing on these support networks. FOSS technologies and secondary support networks have generally not overcome this illicit network and its lock-in effects in many countries—in fact some studies have suggested that, in underdeveloped markets, toleration of piracy is a rationale business strategy for major commercial vendors because it favors industry-leader lock-in and undermines FOSS competition./6/

Another factor shaping cost is the “Microsoft Effect”: a set of Microsoft policies and strategies geared towards preserving and extending their market dominance, especially in the developing world where technology markets are growing rapidly. Microsoft has begun to take measures to adapt both to Linux’s inroads and to the disparities in its own pricing relative to local incomes. The systematic underbidding of large institutional, educational, and municipal software contracts is one strategy. The recently unveiled ‘Windows XP Starter Edition’—a streamlined, cheaper version of their signature OS—is another, destined for contested developing markets such as Thailand, Malaysia, Indonesia, and Russia./7/ Microsoft educational programs designed to provide Microsoft OS’s for donated PCs operate in 67 countries. All are recent responses to the rise of FOSS activity—and at least officially, high levels of piracy—in developing countries.

Many source camp participants are unaware of these programs, but they make and remake software decisions in contexts shaped by the broader market conditions they affect. In the developing world, the lock-in of proprietary industry standards is no longer just a side-effect of piracy. It is actively promoted by increasingly engaged commercial vendors. Because costs of FOSS adoption have much to do with the penalty of working outside prevailing network effects—in terms of transition, retraining, and support costs—source camp advocates are careful to convey the complexity of this terrain./8/

Resituating FOSS: Ethical Frameworks and Conceptual Maps

For source camp facilitators, the most common challenges in presenting FOSS are not misconceptions on the part of participants, but “no conceptions”—no position on FOSS or sense of its possible role. Source camps consequently devote time to framing this relationship, including creating new vocabularies and knowledge tied to the issues and concerns that arise in efforts to bring FOSS and NGO worlds together. Crucial to this process is the effort to align FOSS's philosophy of transparency and accessibility with the broader goals of many NGO institutions to strengthen civil society through forms of community involvement and collaboration. The explicit structure of the source camps, “participatory, non-hierarchical, collaborative, and hands-on,” reflects values that TTC facilitators and participants identified as common to both FOSS and NGO communities. As NGO participants learn about FOSS, many begin to understand it in relation to the values that animate their professional work. FOSS’s novel legal arrangements, such as ‘copyleft,’ also provide an important point of engagement. Because NGOs increasingly perceive global intellectual property regulations as contributing to forms of social and economic inequality, the egalitarian principles of access and dissemination mandated in the copyleft are very attractive.

At source camps, this ethical framing facilitates a number of more practical exercises in knowledge sharing among the participants. Whereas many NGO-sector participants lack adequate understanding of the technical, social, and legal intricacies of FOSS, those involved with FOSS development or advocacy often have little experience of the economic and technological challenges facing NGOs in the developing world. The common values of transparency, egalitarianism, local capacity building, and access to knowledge create a space where mutual understanding and knowledge sharing can emerge. Intermediaries such as eRiders—a class of technology consultants specialized in the technological needs of non-profits—also play important roles in this process. eRiders often have a broad perspective on developing-world information needs derived from one-on-one experiences with NGOs. For these reasons, TTC goes to considerable lengths to invite eRider participation in source camp programs./9/

Knowledge transfer also involves the creation of “conceptual maps” of the FOSS landscape in order to better understand the ways in which FOSS is able or unable to meet NGO needs. These are derived from conversation and formal exercises, and later transformed into needs assessment reports that can be accessed by all parties. Replicated in every source camp, TTC has generated a number of highly detailed maps that speak to region-specific needs, as well as addressing issues, problems, opportunities, and concerns that cut across geographical regions./10/

In addition to rendering the FOSS landscape visible to NGOs, these maps help define IT development and advocacy roadmaps for the future. By visually and conceptually rendering the gaps in availability (for example lack of documentation for particular applications, or the need to train local technology consultants), these maps shape the evolution of TTC’s activities and help define its narrower interventions. As a consequence of these exercises, TTC has begun to place less emphasis on informational activities and needs assessment and more on hands-on technical training. Other notable outcomes include a more aggressive focus on localization, a Migration guide, the Tajikistan FOSS initiative for schools, a Content Management System project in South Africa, and a series of FOSS meetings in Georgia, Central Asia, and Brazil.

Pragmatic Justifications

Despite success stories about NGO migration to FOSS, the terrain of adoption is still bumpy. The politics of open source adoption in the NGO sector continues to resemble evangelism, rooted in efforts to convince NGO of the benefits of FOSS beyond the variables of cost and ease of use. Although knowledge transfer, mapping exercises, and the identification of shared values between NGOs and FOSS inform this process, they are not pragmatic justifications in and of themselves. And although NGOs are often attracted by the potential of FOSS to shift the broader politics of intellectual property, this is rarely a motive that will weigh heavily against practical needs. TTC generally emphasizes three more pragmatic reasons for adoption: security, customization, and localization.

Security

The security claims for open source code are rooted in the argument that ongoing peer review by a large community produces fewer exploitable flaws than the “security through obscurity” approach of proprietary software. The poor security record of many Microsoft products, especially, has brought this argument to a much wider public than was true even three years ago.

In source camps, TTC facilitators make much of the security advantages of FOSS. Not everyone finds this compelling, in part because many participants hold a narrow view of software security relating primarily to the privacy of records. As one TTC consultant explained, the security message “resonates with about 15-20%” of the NGO participants—generally those involved in political activism and human rights, for whom the confidentiality of internal documents or data may have life-or-death implications. The majority of NGO participants tend to view their records as insufficiently sensitive to warrant careful attention to ‘intangible’ differences in security. Often, the vulnerability of systems to attack from viruses and other is perceived as a separate problem.

The availability of tech support makes a difference a big difference in this context, both because system administrators are attuned to questions of system security, and because updating security patches on Linux still requires a degree of technical skill greater than that associated with updating Windows.

Customization and Localization

Customization is intrinsic to FOSS software and to the FOSS developer ethos. The availability of the source code provides complete, if often technically demanding, control over design and features at a much more granular level than the configuration and preference options of most proprietary software. Because of these technical demands, however, low-level customization has usually mattered much more to the developer community than to the end user, and in practice has resulted in a proliferation of versions that can be confusing to those end users. Linux distributions, for example, are so well-stocked with applications (4 chat programs, 3 word processors, 2 accounting programs, etc.) that knowing what to use and trust can become a laborious and frustrating research exercise in navigating help programs and testing software. To minimized this source of confusion, several initiatives are emphasizing streamlined distributions that provide only the “essential” tools to meet the needs of NGOs (Debian Non-Profit and TTC's NGO in a Box are two notable examples)./11/

Customization in the sense of ‘localization,’ however, is becoming more important to the NGO sector. Localization is the process of adapting software so that it conforms to some locally defined need, commonly language. In this case, localization involves rewriting software so that words in menus, dialogs, and dictionaries appear in the target language, a process that often involves complicated character encoding methods for non-western character sets. Because the legal and technological characteristics of FOSS allow and even encourage customization, FOSS is at a marked advantage for meeting the requirements of small language groups./12/ Although FOSS language localization is far from complete, it is in many respects ahead of Microsoft and other industry standards in this area.

TTC has begun to view localization as one of the best opportunities for making FOSS a leading choice for NGOs. Working in collaboration with Aspiration, another NGO focused on FOSS advocacy, TTC recently hosted a “Localization Sprint” in Warsaw (November 2004). A more specialized event than the source camps, the Sprint brought together localization experts and project leaders to share experiences and compare projects in the hopes of creating new collaborations and clarifying best practices./13/

Localization has a number of limitations, however, including its lesser impact in regions dominated by major languages, such as Latin America. The market opportunities provided by major languages are strong enough to ensure that proprietary software will be well represented. Second, the lack of a developed IT sector in many localities undermines the volunteer-based structure of participation in FOSS projects, accentuating the difficulty of funding and sustaining such projects. As project leaders at the recent Brazilian FOSS conference made clear, donor and state funding for FOSS localization is relatively rare.

The Broader Economic Landscape

The precarious financial status of much of the NGO sector and the underdevelopment of the IT sector in many parts of the world shapes FOSS adoption in ways that go beyond calculations of cost. In an era of government cutbacks in funding, NGOs rely increasingly on corporations for fiscal support, including technology grants and second-hand technology giveaways. Technology company philanthropies, especially, tend to make gifts in kind—either in equipment or in services. These often come with or rely on proprietary software, raising the relative initial cost of FOSS solutions or FOSS migration. A significant number of IT solution providers for NGOs, moreover, effectively operate as Microsoft shops—CompuMentor is a prominent U.S. example. When large contracts are at stake, Microsoft has increasingly offered significant discounts. Microsoft’s recent deal with SchoolNet—a large, pan-African NGO that works to improve education through the use of information and communication technologies (ICTs)—is an example.

These philanthropies, recycling networks, and discount programs play an exaggerated role in countries where the private sector in IT services (and consequently in FOSS services) is underdeveloped. As many TTC facilitators observed, NGOs migration to FOSS is difficult without support from a local private sector in FOSS technologies, which can bring paid or volunteer personnel, knowledge, and tech support to bear. Although nearly all of this expertise and documentation is also available on the Internet, and while the developer community is generally comfortable with this medium of support, NGOs generally typically have much more difficulty making effective use of virtual help and documentation, especially when they lack dedicated IT staff. This dynamic is most problematic in the context of migration to FOSS, when end user familiarity and expertise are likely to be lowest. Inconsistent access to the Internet adds to this burden.

To date, local FOSS communities and service sectors are almost always offshoots of (and dependent on) larger proprietary IT sectors, which provide most of the employment opportunities for IT specialists. As a consequence, the great majority of FOSS experts in the developing world make their living off of proprietary software (Microsoft, Oracle, Sun, and other products), not FOSS software. The precariousness of this labor market in most developing countries has the important effect of diminishing the leisure time available for FOSS projects or for participation in local “user group associations.” The time/employment dilemma is cyclical insofar as these are the venues where many FOSS developers acquire and expand their skills. Such disparities increase the importance of private-sector FOSS adopters, including anyone from an ISP that runs on primarily FOSS technologies to a branch office of Hewlett Packard providing Linux support. These become crucial reservoirs of local FOSS expertise. State, educational, and municipal adoption, while capable of quickly scaling up the number of FOSS users, are also very vulnerable to this dynamic.

As the pool of IT workers with FOSS expertise grows—either through direct employment in FOSS services or as a side interest—chances increase of building and sustaining the social networks characteristic of strong FOSS communities. Volunteer associations such as Linux User Groups (LUGs) often play a large role, especially as sources of free or low-cost tech support. These can often be tapped by NGOs./14/ Affordable professional consulting services also tend to grow out of this mix. Although TTC does little to shape these larger sectoral factors, it recognizes that the strongest opportunities exist where the macro-level forces underpinning IT sectoral development (pro-technology public policy, economic growth, high educational capacity) are aligned. These set the stage for lower cost migrations to open source.

The Developer Community

In recent years, FOSS developers have organized a number of programs to provide support for local FOSS adoption in the developing world, notably GeekCorps, Free Geek, and the South Africa Localization Project. FOSS evangelism runs strong in the developer community, and has produced an active fringe of NGOs and other organizations devoted to the egalitarian social dimension of FOSS. Developer-driven approaches have both strengths and limitations, however. Although FOSS technologies may be well suited for NGOs in many circumstances, developers themselves may not be the best actors to make the case. As one TTC consultant with years of experience among hackers and Silicon Valley entrepreneurs explained, developers often operate in “silos,” cocooned in self-referential exercises without much awareness of external needs. Because the majority of developers work on free software for personal reasons—including recognition by their technically talented peers—there has been less focus, until recently, on understanding the usability needs of non-technical users. For these reasons, Tactical Technologies has shifted its focus over time from developers to technology consultants and other intermediaries, such as eRiders and NGO technical personnel. These provide a different and often more accessible version of FOSS evangelism, rooted closer to NGO experience and the challenges of building common ground.

This situation has improved significantly with the entrance of large firms like IBM into the FOSS market, and with the rise of other intermediaries and service providers who target mainstream computer users. Though many FOSS tools remain challenging to install or use, there have been steady improvements in usability in many of the major software packages. Programs like GNOME, KDE, OpenOffice, and Mozilla have altered the landscape of FOSS by providing a high-quality user-friendly experience—equaling and in some instances surpassing the major proprietary systems (the recent success of the Firefox browser is a widely-noted example). Where the installation of GNU/Linux was once a grueling ritual of initiation that tested even experienced users’ patience and skills, there are now a number of distributions more accessible to novice users, and a number of professional developer communities committed to standardized updates and service cycles (e.g, Ubuntu Linux, which draws on Debian, one of the most popular distributions in the engineering community). Many FOSS advocates believe that this hybrid volunteer/business model will open much wider paths to adoption. For TTC and other technology advocates committed to both FOSS as both a technological and social project, these efforts are positively affecting the cost-benefit analyses made by NGOs as they consider FOSS solutions. In particular, they are making it easier for NGO’s to extend their social values into their technological practices. TTC and others are working to help this process.


Footnotes

1 For a central repository of such documents, see Free Software / Open Source Software and Civil Society Organizations - the guide http://fossforum.tacticaltech.org/

2 For example, a good portion of Greenpeace servers worldwide run GNU/Linux; SchoolNet Namibia relies on open source operating systems, email-clients and office applications to provide Internet access and training to the nations' schools. Other examples are in India: Goa Schools Computers Project (GSCP), Sarai Cybermohalla, Gram Chitra, Ganesha Project, Sakura Project. In Africa, Schoolnet Namibia and Shuttleworth Foundation TuxLabs. And in Latin America Nodo Tau in Argentina, the Rigoberta Menchu Tum Foundation in Guatemala.

3 Most NGO/NPs use only three kinds of software: Internet (servers, mail exchange software), Intranet (local network), and Desktop. TTC experience suggests that the majority of FOSS software currently deployed in these organizations is server-side, such as the market-dominating Apache web server software. Desktop and Intranet deployment is far less frequent though becoming more common, with applications like Mozilla and OpenOffice leading the way.

4 This is not by any means unique to the NGO sector. Elsewhere I have argued that one of the defining political elements of FOSS is its ability to act as an “iconic tactic” by which different social groups take the underlying philosophy of freedom animating FOSS and translate it new terms to realize other political or economic goals (Coleman and Hill 2004).

5 TTC is acutely aware of the monetary and human resources required to adopt FOSS solutions, and consequently avoids treating FOSS as a “magic bullet” that can easily satisfy all of the technical requirements of NGOs.

6 See, e.g., Osorio (2002).

7 There is very little serious or systematic analysis of this subject. However, MS policies in the developing world, especially as related to Linux are regularly reported in the media. See, for example Kanellos (2004a, 2004b).

8 The question of cost is also region-dependent. A number of empirical studies have been completed or are underway which seek to assess this question with more rigor and detail than currently exists. See for example, Comparison Study of Open Source and Proprietary Software in an African Context: Implementation and Policy-making to Optimise Community Access to ICT available at http://www.bridges.org/software_comparison/about.html. Also informative is the The LINC Project Guide to Choosing an Operating System available at http://www.lincproject.org/toolkit/cos_guide/.

9 To learn more about this initiative and about eRiding see About eRiding available at http://www.tacticaltech.org/eriding.

10 Several of these conceptual maps are available in report form. For a polished and in-depth report on the state of FOSS in Africa and its potential for NGOs see Straight from the Source: Perspectives for the African Free and Open Source Software Movement (available at http://www.tacticaltech.org/files/straight_from_the_source_may04.pdf). A less-developed version for Latin America, Projects Mapping and Needs Assessment, is available at http://www.tacticaltech.org/node/136).

11 http://www.tacticaltech.org/ngoinabox, http://www.debian.org/devel/debian-nonprofit/

12 See, e.g., translate.org.za, which is working to adapt some of the most popular FOSS applications (OpenOffice and Mozilla) to all of South Africa's 11 languages.

13 http://localisationdev.org/

14 Although such communities track private and public sector interest, there are few strong national or regional patterns. Some cities, like Buenos Aires, Porto Alegre, and Banglore have large and active large associations while in many other regions they are tiny or non-existent.


Works Cited

Coleman, Biella and Mako Hill
2004 How Free Became Open and Everything Else Under the Sun. M/C: A Journal of Media and Culture, 7, http://www.media-culture.org.au/0406/02_Coleman-Hill.php

Ghosh, Rishab
2003 “License Fees and GDP per Capita: The Case for Open Source in Developing Countries” First Monday (8)12 http://www.firstmonday.org/issues/issue8_12/ghosh/index.html#g3

Kanellos, Michael
2004a “Microsoft marches on Moscow to do battle with Linux” Cnet News http://news.zdnet.co.uk/software/windows/0,39020396,39168056,00.htm

2004b “Microsoft picks five countries for cheap Windows” CNET News
http://news.zdnet.co.uk/software/windows/0,39020396,39163137,00.htm

Osorio, Carlos
2002 “A Contribution to the Understanding of Illegal Copying of Software: Empirical and Analytical Evidence against Conventional Wisdom” MIT Working Paper; http://opensource.mit.edu/papers/osorio.pdf .

Peizer, Jonathan
2003 “Realizing The Promise of Open Source in the Non-Profit Sector” September, 2003 http://www.soros.org/initiatives/information/articles_publications/articles/realizing_20030903

Zuckerman, Ethan
2003 “Free Beer Does not Sell” LinuxJournal http://www.linuxjournal.com/article/6785