Everyday life increasingly depends on IT systems. It is therefore vital that projects implementing the applications be delivered as far as possible on time, within budget and with full functionality, including performance and, especially, security. What’s likely to maximise the chances of project success? In many years of experience of IT-intensive projects, I have found that there are four conditions. If some or all of them are not met, the likelihood of underachievement or outright failure is raised.
The first condition is that there must be a clear and agreed objective – what the project is intended to deliver. In addition, all those involved should want the project to succeed. This includes not just the people doing the work, and their management, but also those who are expected to use the result. For example, a project to deliver a system for people providing healthcare should get buy-in from the doctors and others who are expected to use it. Buy-in is essential for professional-level users such as doctors: if they see no value, they will subvert the system or simply not use it.
The second is that there should be sufficient resources available for implementing the project, and during its life after completion. Resources include time, and money to pay for the people, facilities and products required. In this context, products would include development tools and methodologies, and the hardware/firmware and software platform in which the applications will run. The choice of platform in particular will contribute heavily to reliability, availability, security and performance. ClearPath Forward systems, for example, provide a sound foundation for critical applications, given their high levels of performance and security.
The third is that the people doing the work should know what they are doing. Expertise and experience really do matter; methodologies and development tools do not compensate for lack of expertise. And expertise is necessary for all the people involved, not just technicians. Those making purchasing decisions, for instance, should understand what they are buying or they are likely to succumb to the lowest cost shoot-out.
The final condition is that there should be a willingness to report problems and to deal with them as quickly as possible. There is always a danger that problems are swept under the carpet because they are inconvenient. Those pointing out difficulties are sometimes discouraged from doing so and even penalised. The right approach is to encourage people to raise doubts and questions at the earliest opportunity. The only offence should be failure to report difficulties.
All this seems to me to be obvious, so it’s surprising that projects start and continue without meeting the conditions. In particular, the final condition – responsiveness to problems – is all too often ignored in IT projects. The result is that the industry does not learn from its mistakes. This is in stark contrast with civil aviation, where every deviation from the expected, up to and including accidents, is investigated and lessons shared across the industry.
However, when all the conditions are met, what can be achieved is spectacular. The Large Hadron Collider (LHC) particle accelerator at CERN is the world’s largest machine. It is the result of one the most complicated projects ever undertaken. CERN is the Conseil Européen pour la Recherche Nucléaire. As well as fundamental particle physics research, there have been many technical spin-offs from CERN; the invention of the World-Wide Web is perhaps the best-known example. (See http://home.cern for a wealth of information about the organisation, its history and activities. Professor Lisa Randall of Harvard University devotes a lot of space to the LHC in her excellent book ‘Knocking on Heaven’s Door’; it’s well worth a read.)
The LHC, with the detectors used to gather data from the collisions, are among the most complex pieces of technology ever built. They are right at the limit of what is technically possible. Indeed, it’s amazing that they work at all! So far, the results have contributed significantly to our understanding of nature, including the discovery of the Higgs Boson in 2012.
Thousands of people in many organisations have been involved in the LHC project. The IT components alone were massive projects, critical in the design, implementation and operation of the LHC, and in the analysis of the data gathered. The Compact Muon Solenoid (CMS) detector, for example, generates about one terabyte of data per second. Only a small fraction – still a lot of data – is saved and processed in the Worldwide LHC Computing Grid.
While there are many complex IT-intensive projects, few approach the scale of the LHC. It seems reasonable to assume, therefore, that if the focus on the conditions for success that characterised the LHC were applied to other projects, we would be much more confident of getting the quality of IT applications on which we depend. The LHC project shows what can be achieved.