My academic background in agricultural sciences, combined with information technology and the innate desire to contribute to the improvement of the lives of people living below the poverty line in Africa, have influenced the direction of this work.

In this thesis, the problem was how to design and integrate interactive information architecture into an information ecology. This is because without integrating technology carefully into the context for which it was designed, the most frequent consequence is that it would fail to be put into use. Nardi and O'day have clearly recognised the magnitude of this problem, in working with the metaphor of information ecology and suggesting the involvement of mediators. Mediators who are keystone species can adapt the tool to fit the local circumstances for which it is designed.

Davenport and Prusak stress that information architecture that serves to guide the user to the location of information enhances the findability and usability of such information. This is because you cannot use what you cannot find.

Interestingly, new technology might create a sensational entry into the information ecology through extensive marketing, but is soon lost without being integrated due to the fact that it has not been shaped to fit the use situation. For instance, it could be too complicated for the human subjects to learn to use, too expensive for them to maintain, or could just be a poor repetition of what they already have. Moville and Rosenfeld stress the notion of a holistic information architecture which results to a solution that is workable within the information ecology given.

The methods used to find out how to design the information architecture known as ”M-MKulima” in this master thesis were interview, benchmarking, background research, a heuristic evaluation, persona, use cases, and prototype testing. The interview gathered information and data that helped establish the current information practice, barriers in the current practice, who the provider of information was and the keystone species. Benchmarking supported the adoption of good ideas that were working successfully from similar information architectures while a background research was used for a technology assessment to establish what kind of technological tools and functions could be creatively built upon. A heuristic evaluation was suitable for identification of the existing structures in the ecology in order to build on what was already working, while persona and use case methods were used to help focus design decisions around meeting user needs as best as possible and for specifying system requirements. Finally, a prototype testing was to pinpoint the weak and strong usability and information architecture features of M-Mkulima. The results enabled the weak features to be improved upon so that M-Mkulima could function more effectively, saving the cost and time of finding information.

M-Mkulima is an emergency system to help farmers manage an endemic plant disease known as Maize Streak Virus. In its integration, it was established that the agricultural officer was actually a key stone species, and thus a human mediator. In line with Nardi and O'day's explanation, and the Schema theory, this meant that he could facilitate the integration of the information architecture successfully into the ecology, because he had a specific role within the organisational environment. This role meant that he possessed some very necessary skills and resources.

Consequently, the problem formulated, which had been how to design and integrate information architecture into an ecology was answered. This was through the information architecture development process of M-Mkulima and the identification of a mediator through whom careful integration into the ecology could be facilitated.

This work therefore supported and expanded on Nardi and O'days work, which augments the use of a mediator in integrating new technological tools into an ecology.

In order to improve the proposed concept of integration and design and get fresh perspectives, I have included formative evaluation and test of concept outcomes of qualified professionals within the agricultural sciences field as well as information technology in Kenyan and Denmark. I have also attached a research paper where ”Saidia Mtoto,” a mobile emergency system prototype whose concept was the basis for working with M-Mkulima, was reviewed and accepted by experts at the 3rd Workshop on Innovative Mobile Technology and Services for Developing Countries (IMTS-DC'10). The workshop was in conjunction with the International Conference on Computing and ICT Research. (ICCIR10), Makerere University, Kampala, Uganda.