A response to this line of Dan Sheldon's, in my brisk manner.
For your readers, the question they should be at the stage of asking. Is, so what? Whilst I absolutely agree that there is a huge, huge mess of stuff going on in some parts of UK Government, it took more than 35 years for some of that mess to get there. It’s a key part of lean-agility to make incremental changes, and since you can’t go faster than the culture let’s you change (big programmes find that out the hard way) changing that level of technica legacy, which has 35 years worth of knowledge about the domain baked into it, from people who have not only retired, but have died (if the DWP tech talk in Manchester was anything to go by) is not an overnight activity as that is just a Big-Bang waterfall transformation that, like all Big Bang transformations, is subject to higher levels of failure than the equivalent smaller set of transforms. That statement is both empirically evident and one borne out in mathematical proof (it genuinely is a stochastic exercise). To quote an earlier point I read, you can’t ban maths. So you have to work with that knowledge and the richness it brings, to drive your programmes with a statistically higher chance of delivering value each time.