UK government's newest department to lead mega ERP procurement A five-month-old UK government department is set to lead a massively complex ERP procurement to bring together software running some of Whitehall's largest units. Expected to bring together up to 48,000 users on 15 instances of nine different software solutions, the tender for the Matrix cluster of the ERP program was …
My suspicion is that legacy systems - and more importantly - legacy bespoke processes will be retained, leading to an integration nightmare rather than a standard agreed deployment across different stakeholders.
This will lead to a constant tinkering on the integration side, an industry around consultancy and implementation specialists, delays, additional cost and in the end a system that is just as unwieldy as the separate systems were, leading to a total lock-in.
There should be a target architecture with standard processes and interfaces, meaning that at some point in the future a potential migration may be a large undertaking, but far less complicated than going through the same mess that will be waded through during the implementation phase here.
Wouldn't it be nice to actually learn from the past and plan for the future?
Don't these people ever learn? ERP is a hellscape at the best of times but lob in assorted goverment departments and it's a full on 100% crash and burn and that's before you start tagging everything with restricted markings, hiding embarassing stuff and selecting approved (and obfuscated) users.
I don't know who is in charge of the Dept of New Stuff but I do know s/he is out of their depth already.
.. but that would involve rolling back some of the stupidity infesting government IT (it involves applying some techniques that are normally used to speed up M&A integrations).
Given that the aforementioned stupidity is *extremely* profitable for some I don't expect to see that happen any time soon.
I'm assuming there is some sort of roadmap to achieve this, it'll be unfunded as usual, i.e. no actual idea of the amount it will cost. One of the consultancies will come in with a boilerplate approach, charge millions, then bleat they weren't there to deliver anything but just advise.
Work with universities!? To develop software?! Are you having a f**king laugh?? Have you ever seen the quality of code that comes out of most universities? For part of my Master's CompSci project I had to try to extend a million line Java monolith codebase written by some "genius" with zero documentation, zero use of established patterns and practices, zero tests.
Have you seen the infamous codebase that was used for the COVID modelling? It was terribly written. And there are plenty more examples where they came from.
And that's before you take into consideration that most academics reside in ivory towers with no idea of how the real world works.
Honestly getting Universities to build software might be cheaper than consultancies but the chance of getting anything even close to fit-for-purpose and production-ready is even slimmer than if the usual suspects are paid through the nose for it.
2019 Decide to take IT back in house.
Switch from heaviliy bespoked SAP to Oracle.
Estimate £20m
Current estimate £100m
Migrating a single major ERP to another single ERP.
This top-down centrally planned migrate-and-merge has f**ked up so many times before what's the odds HMG wil succeed? 100:1?
Don't get me wrong. Data centres should be merged to the places with the lowest operating costs. Remote working staff should not have to migrate if that's somewhere remote (next to a hydrolectric dam in Cumbria for example).
But it should be organic, encouraging mergers, not forcing them.
Of course following the successful campaign by SI's to emasculate the Civil Service of not just people who can do technology but even people who can manage the process of doing technology they have no f**king clue of what a s**tstorm this will be.
…. so after 10 years someone will have had a “better” idea and some other vendor will be invited to have a crack.
Far better, as discussed above, to migrate each department to the same core IT so they could (maybe) be integrated in the future. But just integrating their Active Directories gives me shivers down my spine, let alone anything else!
Completely mad. Agree standards to remove interdependencies and let them use what they have (skills and investments) where they are. It’s a classic System Integrator money binge fest that will not deliver a bean other than years of heat.
The best thing we ever did was to implement a system against a very tight timescale, which by its very nature relegated all the “we’ve always done it this way” type perversions that lead to unnecessary customisations. Build out once you have the code. There is a better way.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
