Here’s a couple of stories from today’s press that tell you an awful lot about why getting the requirements right in the first place – and then making sure that those requirements are what the end-users really need – are pre-requisites for a success project. They also tell you all you need to know about large-scale government IT projects:
Daily Telegraph: £1.9 bn portable radio system gets a poor reception from Army
The Army’s new £1.9 billion communications system is facing such serious problems that the Director of Infantry initially refused to accept the portable radio into service.
This story puts me in mind of the oft-quoted example from requirements specifications – where the customer asked for a swing, and the contractor eventually delivered a bird-table. Whereas the old radios ran on AA batteries, the new ones use bespoke (and irreplaceable) ones – and is tied to the operator by straps that can’t be easily undone if they’re incapacitated.
Then there’s this about the new all-singing, all-dancing MOD computer system:
Daily Telegraph: MoD insists on computer meltdown clause
The aim of the new project is to create a system providing everything from email services for soldiers to ammunition-ordering from the battlefield. It will comprise about 140,000 computer terminals and kiosks.
(Emphasis mine) And what are the chances of a system doing everything from email services to ammunition-ordering doing any of these things well? It seems that the concept of small pieces, loosely joined – or even doing a few things well rather than all things badly – is one that’s usually lacking in large projects. What starts off as a simple system to do X rapidly becomes loaded with features Y and Z as well – with the end-result that unintended interactions between all three parts causes the system as a whole to break down.
No doubt the new radios were intended to be a significant improvement on the old ones, but it’s doubtful whether the intended benefits will be gained if the story is to be believed. Instead the replacement system will need extensive reworking to meet the operating requirements, and end up late and over-budget as a result.
