Designing Software for a Million Users: it's not About the Technology

Scotland was probably the first country in the world to deploy a national digital learning environment in 2006. This was an innovative and pioneering system but suffered from two fatal flaws. There were problems with the system’s usability and it was a closed system. It offered a set of applications that were designed in the early years of the 21 century. Much more effective and usable applications had become freely available on the Internet before the system was deployed. Our goal was to specify a replacement for this original system that avoided the flaws in its design and which provided a range of opportunities for teachers to use digital technologies to support learning in their classrooms. We wanted to create an open system that could integrate the web services and applications that teachers and students really wanted to use. Potentially, this system has a large user base. There are more than 3000 schools in which it will be deployed ranging from tiny rural schools with a handful of pupils and a couple of teachers to huge city school campuses, with thousands of students. There are more than 50,000 teachers and almost potential student users. As an aim was to use the system to involve parents in their childrens’ education, there are more than a million potential users of the system. The key challenges that we faced were not technical problems. Issues we had to address were a very wide range of user abilities, complex system governance, a hostile user base, heterogeneous hardware and a desire to avoid political risk in deploying the system. For many types of user-facing system, the technology is not the problem – it’s the socio-technical stuff that causes the real difficulties. Socio-technical factors are the reason why many technically excellent systems don’t really work that well. They’re the reason why so many software-intensive systems projects are late, over-budget and fail to realise their objectives. As engineers, I think it’s time that we faced up to these issues and thought about how we can better understand how they affect the systems that we are creating. Proceedings of STPIS'15