RDVP Seminar: Steve Mushero (2/2/2005)

Steve Mushero came to address the fellows about the topics of developing economies and having NGO's as customers. As a global technologist who creates systems for things like disease tracking, Steve has experienced first hand the challenges that many of us are facing or will be facing as we move from the planning phase to the execution phase. Many of those complexities come from the messy details of the real world that don't make it into the abstract systems that we design for. Just to mention a few, Steve talked about how wars impact Geographic Information Systems (GIS):

• Disputed boundaries
  
• Overlapping governments
  
• Unclear authority of who owns the data or gives authorization to release
  
• Refugee populations
  
• Towns that move to safer locations
  

Working with NGO's can be hard as well, as they often:

• Have different expectations than the business world
  
• Expect alignment at the mission level
  
• Are composed of volunteers who may have more passion than experience
  

Regarding specifically the technology infrastructure:

• Managing support for people far away, from a different culture, who speak a different language
  
• Insecure computing environments where viruses and spyware are prevalent, with fewer trained system administrators to manage the machines
  
• Harsh operating conditions for hardware (intermittent power, extreme temperatures, sand) and difficulty in getting replacement parts
  

Steve advocated a very participatory design and prototyping process. Users don't know what they want or what's possible until they see it (or a near miss). But it's hard to design a system to do a job until you can do it yourself. "Miracles happen when you send engineers out to actually do the job," he said. He also argued for maintaining simplicity and reliability, striving to be at the sweet spot of the 80/20 rule, getting 80% of the benefit for just 20% of the total system work. Defensive programming (including diagnostics and startup tests) can help to reduce work in the long run.