Big IT and integration
“this is about management rather than technology”
What does this refer to? It’s apposite to so many projects (for instance the failure of Freiburg to adopt open source systems), but this instance is about the DWP’s nightmares getting the new Universal Credit systems running smoothly, as detailed by Rory Cellan-Jones on the BBC this morning.
Why do big IT projects struggle?
Let’s not underestimate the scale of the task: creating the systems to run the new Universal Credit was always going to be a vast and complex job. However, systems have to be up the tasks required of them and to have been designed both for the material they will handle and for the people that will use them. Rather briefing against IDS, a civil service insider had this to say about the tremendous efforts being made by the staff at the frontline, referring to them:
“struggling so much with the number of times they have to re-key, systems are crashing. They’re not joined up, they just can’t cope with the messy reality of people’s lives”.
This project is struggling in part because it’s using a system heavily criticised by the National Audit Office. It’s proprietary, developed by Accenture and IBM. Why do I mention that it’s proprietary? Because among other things that means that it hasn’t been open to critical review and only the original authors can amend the code, and therefore supporting it is much harder work and ramps up the cost of ownership enormously. As Shameem Hameed says in this article on IT in healthcare:
With proprietary software, the amount of developer resources that can put into refinements may be limited to that one vendor’s resources. With open source software, countless companies and individuals are constantly collaborating to make the software easier to operate and more user-friendly for everyone.
Would open source benefit Big IT?
It would. Depending on how far you want to open source a project – ie all the processes that go into informing the design, just the source code or somewhere in-between – every element is open to peer review. Dodgy code is not obscured behind a veil of proprietary secrecy, but identified and corrected. No-one would claim that open source code is always perfect, but the openness means that problems are identified early on and a large body of expertise can apply itself to the problem.
Imagine if that had happened in this case? Would £300m have been written off? I don’t think so.