Is it really possible to bring together on a single Windows-based infrastructure hundreds of varied systems which operate at different levels of security at hundreds of defence sites at all three services, the army, navy and air force? We have our doubts.
If it all works it could save huge sums being spent on perpetuating a variety of systems. It could also help troops at, or near, the front line where information is presented to them from different systems that do not always talk to each other.
But £5bn is a lot of money to risk on a project that may or may not work. And disruption at MoD sites may sometimes be too serious to be trivialised as the inevitable result of a new system bedding down.
Computer Weekly and Channel Four News has learned, as part of our joint investigation, that the rollout of the DII project to the Infantry Guided Weapons integrated project team may significantly affect the work they do.
It is likely these problems will be resolved in time. Or they could be symptomatic of a project that, in its original concept, is too complicated, too ambitious. A great idea does not justify an impractical proposition.
The DII cannot fail completely - 16,000 systems have been delivered. But the plan is for 150,000 standardised terminals to be installed across 2,000 MoD sites worldwide. That looks to us to be a grand hope. It may not be possible given the current difficulties and shortages of money.
Installing new hardware is not always difficult - the big challenge is to migrate data accurately and reliably and run tried and tested systems on the new infrastructure.
This is a huge problem even without the challenges of trying to standardise on applications that work under different security levels. And then there are difficulties such as cabling in barracks, under runways, and in organisations that seem routinely to change shape and size. Roles and locations of thousands of personnel also change often.
We are struck by the professionalism and commitment of the project teams on both sides: the MoD and the Atlas consortium led by EDS. But are they working on mission impossible?
They will not want to give up on the DII and we do not think they should. There have attempts to make the scheme simpler, to "de-risk" it by breaking it into three increments, and even to phases within these. But there is a danger the "one-size-fits-all" technology could become so simple that it is derided or even avoided by users who will keep to their existing systems - thus leaving them with two computers where they had one before.
Like the NHS IT programme also - NPfIT - the assumptions that underlay the award of the Defence Information Infrastructure contract need, in the light of experiences so far, to be tested by an independent, published review, at least to assess unemotionally what is, and what is not, achievable.
Ministers do not like independent reviews of major IT programmes because investigators are likely to find things that could embarrass them, such as fundamental weaknesses. But there is too much at stake for ministers to put potential embarrassment before concern for the troops who are already under great pressure and the teams that are trying to get a project of labyrinthine complexity on track.