If I understand your question, you are wondering if building a distributed inventory application is the right way to go. The network connections you have to Africa are weak or nonexistent, and workers there need a system sooner rather than later. Is this a fair summary?
Poor Internet connections in developing countries can derail the best networked application plans. I remember a situation in Pakistan a few years ago. We were looking at an HR information system. For a number of reasons, we wanted it to be a central application in headquarters that all could share in the field. So it needed to be web-enabled and work with a browser. The problem was the Internet connection. In the fall of 2001 we got a quote for a DSL line in Islamabad for $3,000 USD per month. That was too expensive, and we were about to shelf the plans for a single, central system. Luckily we waited. A year later a DSL line for double the speed costs $300 USD per month. That’s a ten to one reduction. We bought the line.
This experience has recurred in many locations throughout the world for us. Internet access is changing monthly, and costs continue to drop. It led us to an important principle: don’t bet against the network. By the time it takes you to build a system that works around a poor or nonexistent connection, the connection you need will be there. The time and dollars are better spent shoring up the network.
So when you are tempted to take the short-term fix, wait. Talk to local and regional Internet Service Providers. Find out when the network will be where you want it to be. Two final thoughts: building a good distributed application always takes longer than you think. Second, when you finish the work-around project you may be in the awkward position of needing to throw it away. |