A bridges.org article in collaboration with the Tactical Technology Collective.

For a software developer working in Africa, Philip Mbogo's problem is as basic as it gets: "I don't have a computer," he said. "I have to go for unpaid work in order just to get on a computer." Internet access is also an expensive rarity, so he counts himself fortunate to work as an intern at an Internet service provider where he takes as much advantage of the bandwidth as he can. "Anything I can get I download. I even got [a Linux distribution called] Debian, which takes two days [to download]."

African software developers face many obstacles as they struggle to work in this field. But these "coders", as a group, form a community marked less by their frustration and isolation than by their perseverance and resolve. This theme dominated AfricaSource, a workshop held in Namibia in March 2004 and organised by the Tactical Technology Collective, AllAfrica Foundation and SchoolNet Namibia. The meeting in the small town of Okahandja of 40 software developers from 25 countries was for many the first chance to collaborate and compare notes.

Lack of access to the means and tools of production is the issue African programmers most commonly identify as the greatest barrier to success in their work. But at this event, coders got a chance to share the innovative ways they work around the problem. "We buy computer parts bit by bit. In the space of three or four months we have a computer," says Ayeni Samuel Olaoluwa, a web developer from Nigeria, who saves up to 50% this way. Another method he has devised is keeping his freelance clients' work on computers he uses as part of his day job. "I am able to hide stuff on the server, but when I leave the company I'm in trouble."

The prohibitive costs of bandwidth and hardware are an obstruction most programmers face, but it affects coders most seriously at the time they are preparing to enter the job market. Without the opportunity to earn salaries that would help them afford equipment of their own, ambitious market entrants eager for work face the prospect of successive, often unpaid, internships just to prove their skills.

This predicament is widespread across Africa, says Ghanaian Guido Sohne, "There are not enough projects available to work on to employ the available talent…. In most African countries IT is not part of the economic production process. It's actually more expensive to computerise your accounting system than to hire more people to do it manually." So when programmers do find jobs, a large percentage tend to find themselves ushered into system administration and technical roles, where they are overworked and their skills are underutilised.

This situation might be a consequence of the fact that a coder's skill is not accorded the value it deserves. Isaiah Makwakwa finds himself in this bind. When the computer science graduate first joined SDNP Malawi, the UN's sustainable development program, his work stimulated him. He automated a billing system for the administration of the .mw domain and maintained the webserver and mail systems. However, over time and as budgets became tighter, managers added client support to his job description. "User support grew to be the biggest part of my job," he says.

The flood of work helping people solve problems on their desktops gradually overshadowed his programming duties, but Makwakwa's manager failed to implement a plan that would have protected and leveraged the value of Makwakwa's skills, which are rarer and far less transferable than the troubleshooting talent he was being called upon to use.

Makwakwa's case is not unique, and it certainly is not exclusive to Africa. As the number of computerised workplaces in Africa has grown, so has the demand for the kinds of skills Makwakwa has gained from his training. The trouble is that he feels he cannot use all of his skills. Many coders, though they are content to be at least employed in countries where joblessness is chronic, still languish in jobs that do not satisfy them.

Ayeni has even faced an anti-African attitude when it came to his work. "I often approach buyers [with my products] but I hear, 'this thing is not developed in America or Europe. We don't want it'". Sometimes being made in Africa does not kill the sale, but it certainly hurts the seller. "They force you to sell it at a cheap price when they know it comes from Nigeria. They underprice you. They don't think it is worth anything."

Market resistance is somewhat stiffer if the services people are trying to sell include free and open source software (FOSS), whose low initial costs and adaptability are showing promise as customised, locally produced solutions for undercapitalized African markets. FOSS is not yet well known enough, developers say. Mbogo says proprietary software like Microsoft remains first in the minds of customers and employers. "They don't ask if you can program. They ask, 'do you know (Microsoft) Visual Basic?'"

Advocacy about African innovation and talent -- especially to educate African decision-makers in government and industry about the merits of using locally developed software -- can go a long way to changing the entrenched beliefs people like Ayeni and Mbogo encounter. "We must start a crusade," proposes Wire Lughabo James, whose East Africa Centre For Open Source in Kampala supports a push for greater recognition and adoption of FOSS, and, by extension, for the African coders who would also benefit from a change in the mindset of potential clients.

But Wire does not envision a crusade fought from within. "Developers all play a good role. But they themselves are not enough to create change…. There is a need for a force that brings together business, policy, strategy, commerce. Does a legislator understand [the encryption tool] GPG? And does he even care? He is concerned only in the policy -- in what explains how process 'X' will benefit society. That's all the President wants."

"One day, though," Wire continues, "the Minister will hear others saying, 'do you hear of what this Ugandan company has done?' And he will not know anything about it. He will be embarrassed. But then he will slam his fist on the table and say 'We must meet these guys tomorrow!' And then we will start the dialogue."

Alongside mainstream industries and big market targets like governments, non-governmental organizations (NGOs) and community service organisations (CSOs) are pegged as beneficiaries for the enthusiasm and passion of software developers. They typically lack the kind of expertise in technology that might help them increase their impact in their sector, and many hope that FOSS developers will be the ones to produce the applications and localise the solutions appropriate for NGOs. But proceedings at AfricaSource showed that there remains a distance to be crossed before NGOs and the FOSS community can come together.

As with government and industry, awareness of the option free and open source software presents users and administrators is only a preliminary barrier to its acceptance among NGOs. "There are many CSOs interested in migrating to open source platforms, but the support isn't there," says Natasha Primo, of WomensNet, a networking organisation. "You don't have a techie in-house… so a lot of CSOs are finding themselves trapped." Manuals are another problem. "The fact that documentation is not always up-to-date also has an impact on CSO's interest in shifting to those applications. It's a nightmare if it's not documented properly."

For their part, developers themselves aren't very well prepared to respond to NGO needs either. "Open source work is done by geeks. It's stuff that they want to use, not some stuff CSOs would need," says coder Francois Bonin of Ghana. Even if a CSO's demands could motivate development, he says, the kind of work that coders excel at fails to suffice for the project to succeed. "NGOs have complex needs -- say, 40 field workers in 2000 villages. It's a business. It's a real business with real problems and they need business-quality applications…. So you need real management."

Removing the impediments to mutually satisfying work between the FOSS community and non-governmental organisations depends on collaboration of agents other than coders, such as skilled project managers, and those who can act as go-between in the absence of a common NGO and software developer vocabulary. "I see the need for an intermediary of some sort," says developer Alaistair Otter. "I'd like to see a true developer's network available to respond to users' needs." But the focus does not have to fall on customisation. "There's quite a scope for providing NGOs and CSOs with education and advice so they minimize risk in going with open source solutions." Most of all, such resources must avoid the pitfalls of developers' portals, whose strongly worded opinions only intimidate people from asking basic questions. "You want to be able to get an answer to a question without starting a [fight]."

The productivity of a maturing free and open source software development community is also beginning to erode some of its potential. Content management systems (CMS) are a classic example, Otter says. The oversupply of CMS options actually discourages people from trying one out, simply because good basic advice about which utility provides the best fit is not available. So consultation about software choice must play a major role in this new network, Otter says, but "there is also scope for [custom] applications. Developers say that this is what they would like to do, but [so far] it doesn't match the needs of the group it's designed for."

Currently the highest-profile strides in FOSS development in Africa are being made in localisation, the practice of translating the language and sometimes the icons of the computer interface into another tongue. Zulu, Afrikaans, and northern Sotho language packs have been released for Mozilla and the productivity suite Open Office by translate.org.za, a South Africa-based organisation. AfricaSource also witnessed the creation of an easily configurable Kiswahili spellchecker for Open Office. That same week, colleagues of Wire Lunghabo James released a version of the Mozilla web browser in Luganda, a language mainly indigenous to Uganda. Its positive reception is, in part, a consequence of an innovation as transparent to novice users as it is to developers. "People were so resigned to the idea that computers were in English," says Wire. "When you tell them that it's in Luganda they cannot believe you," he said.

Language adaptation can also raise problems atypical of the software development world. As programmers without much experience translating, Wire and his colleagues struggled to convey the sense of a word whose meaning in a technical context is so far removed from its conventional use. "Cookie, for example. What is a 'cookie' [in the context of web browsers]?" Wire recalled asking himself. He and his colleagues knew they could not transliterate the English, but they could not immediately find an alternative. "We settled on 'obutonya'," he said. "Things that drop."

Much of the groundwork in finding ways to get the free and open source software community to service the African market lies ahead, but frequent successes are building hope and momentum. As one participant put it, the workshop itself is not enough, but "what's historic is there was a meeting where people came together."