Our approach is founded on the following questions and principles, which we will apply to all government ICT:
Does a policy or a business process need an ICT solution at all? Should the government be readier to say “no”?
If yes, is there an off the shelf product that will deliver most of what’s needed? Or a solution already developed for another part of government that can be adapted? Can the policy or process be adapted to slash the cost of the ICT solution?
Where a new bespoke solution is needed, we will apply a presumption against a total project size of more than £100m, reducing risks to a more manageable scale, and opening up more opportunities for smaller and UK-based suppliers.
Is there an Open Source solution, saving development and licensing costs, and reducing dependence on long-term oligopoly suppliers?
Where any bespoke computer code is written for the government, unless it genuinely pertains to national security, why can’t it be released under open source licences?
Why can’t the front line be allowed to do their own thing, subject to common standards, especially on interoperability?
Why can’t we put citizens in control of their own personal data?
Is there a powerful reason why outside organisations and the public should not have open access to government data?
Why shouldn’t all government ICT contracts, outside the national security arena, be made totally transparent?
Apart from the first principle this is all good, especially the decentralisation of solution development within constraints of interoperability.
The only complaint is that in the 21st Century it doesn’t make sense to imagine that process change will not involve some form of ICT change.
This is also why the budgets for ICT will remain high, no matter how well managed.
Two observations.
First there is now a third alternative to off the shelf and a bespoke traditional code built solution. Business Technology exists that allows custom build where the core code never changes. The emphasis is on the people and the process with business logic that never changes. This allows build of any function large, small, simple or complex. No need to send offshore, no need for taking a risk with inflexible off the shelf or custom code with lack of future agility.
The second is a consequence of the first think maximum project size of £50m not £100m!
Oh yes and it is lead by British innovation at its very best!
Is there a need for the business process in the first place? Cut down on state sponsored delivery?
By seeing it as a customer experience problem rather than an IT problem, enabling contestability in the customer supply chain might obviate a need to get involved in the software decision in the first place.
£100 million buys a lot of bespoke and discourages pilots. Insist on a pilot, with clear and externally agreed _failure_ criteria to avoid the one-last-push approach for projects beyond resuscitation.
There’s always an open source solution. Look at what is happening in the mobile phone market
Use the GPL V3, avoid patent lock-in, never mind Linux, look at Nokia and “Qt” look at the huge ecosystem that is KDE
The front line doesn’t have to be in the public sector in the same way that Nestle don’t build their own supermarkets. And once it’s not in the public sector it will be free to do its own thing.
Insist on open unencumbered standards.
Government must keep out of the way. Not everyone want to shop in Harrods, not everyone thinks it’s bad to shop at Matalan.
Citizens are in control of their personal data. It’s just that government insists on keeping copies of it.
The “Free Our Data” campaign has already done the work on this. In the Free Software world “Marble” is already providing an alternative to Google Maps. Ordnance Survey could cooperate with both rather than try to compete. A different way to save money.
Transparency is a sine qua non for good governance. As they say in Apache (server software) “If it didn’t happen on list, it didn’t happen”
Gerry wrote: £100 million buys a lot of bespoke and discourages pilots. Insist on a pilot, with clear and externally agreed _failure_ criteria to avoid the one-last-push approach for projects beyond resuscitation.
+1
..and £100 million provides multitude places where people who don’t really add anything to the process can hide and push for that last mile.
I’d love to see every procurement described purely as a requirements catalog – as a challenge to the open source community. Not to discourage competition, but provide a reasonable baseline against which project budgets can be compared.