What I never understand with these enormous projects is what proportion is consultancy waste/profit, and what proportion is "genuinely, this is a very large software project".
IT projects that cost billions are incomprehensible to me, outside of frontier work that the AI companies or FAAANG might do. Is the problem really that hard? Assuming building in-house capability is off the table, is there no space in the market for a good software consultancy? Or are they just outcompeted for getting the contract?
> IT projects that cost billions are incomprehensible to me, outside of frontier work
Getting a large bureaucracy to agree on a question where any answer will disadvantage some members of it is frontier work. No-one has a good general solution. Very few people even have approximate solutions for a few special cases.
> is there no space in the market for a good software consultancy? Or are they just outcompeted for getting the contract?
You can't grow big enough to bid on these big projects and stay good.
The people in charge assume that a large scale project requires lots of work to implement in all cases. That is not a given. Creating software that runs 10,000 separate instances is a lot easier than something that has to scale to one instance 10,000 times the size.. There is a bias to 1) assuming the former is what you need and 2) budgeting for the former and 3) preferring the former.
Governments are used to funding a lot of things like transport and the military etc. where costs roughly scale linearly if you are lucky. Usually worse because big projects are so hard to manage and require extra layers of management. If you are really lucky there will be economies of scale.
The people able to bid for big national contracts are also going to be big consultancies who are biased to big and complex projects. For one thing they also have to be able to deploy and support it. maybe if development was separated from deployment and support, it would be better?
There have been multiple examples of things such as smallish government websites (I mean small information sites - the sort of things a freelancer could do) costing millions because its done through a big consultancy.
it's a self-created boondoggle designed to extract money - a scam
that's why the culture created the insane issue of jailing people instead of investigating numbers and fixing bugs - the scam must continue to enrich the players
Don’t forget to mention part of those players are the clients.
“Clients” usually are even worse offenders than these extractive scam companies, mostly because the “clients” are playing with someone else’s money, usually the taxpayer.
The problem for the UK is the Fujitsu owns the last vestiges of the UK IT scene.
ICT/ICL and a bunch of other were all bought out by them in the early 90s.
But also there are two other issues:
1) there aren't many companies that are able to bid on that scale
2) if you remove fujitsu then you're handing it to either crapita or atos. who are also shite.
The solution to this is develop decent in-house architecture team who are able to drive specs and keep an eye on developments.
But
That requires three further things:
1) long term vision, with enough run way/agency to affect change
2) enough money to pay and keep decent people
3) the ability to terminate contracts when they get into shit. This is policitically challenging because contract termination == the current minister is shite in news term. "FLAGSHIP PROGRAMME SCRAPPED BECAUSE OF POOR PERFORMANCE, here is why $government is bad with lots of input from people with agendas that haven't properly been vetted because the journalist isn't a specialist (or its not a news site)"
Part of the answer is dividing programmes up into smaller chunks to attract a wider range of bidders.
The problem is that, even when divided up, they're still often "build and run" contracts which are less attractive to the smaller consultancies which tend to specialise in the "build" part, and are much less keen on the "run" aspect (with good reason, as it tends to be low-margin and painful).
And even then, fashions inevitably change and suddenly someone starts wondering why they've got 30 different tech vendors and need an expensive enterprise architecture team to glue their work together. Cue "why can't we just consolidate all this and get AtoS to do the whole thing?", and the cycle begins again...
> 2) if you remove fujitsu then you're handing it to either crapita or atos. who are also shite.
Yes, agreed, although (at least) Capita are probably much worse. They're currently in a deep hole having screwed over the administration of the Civil Service Pension Scheme. It's been causing deep pain for people for a long time now. Where's IBM when you need them, or Oracle? Just joking.
Fujitsu (well, ICL, the UK company that was bought by Fujitsu and then became the UK branch of Fujitsu) is still the supplier of Microsoft software to the European institutions through a long-standing framework contract with DIGIT.
Why EU could not buy from Microsoft directly to spare the embarrassment is beyond my comprehension.
Why in the world is the ban self-imposed? People literally died because of this, others spent their lifetimes rotting in prison.
A software bug is perfectly fine, everthing has bugs. It was the years of gaslighting, lying and sending people to prison when you could've easily admitted the issue. Ridiculous.
What I never understand with these enormous projects is what proportion is consultancy waste/profit, and what proportion is "genuinely, this is a very large software project".
IT projects that cost billions are incomprehensible to me, outside of frontier work that the AI companies or FAAANG might do. Is the problem really that hard? Assuming building in-house capability is off the table, is there no space in the market for a good software consultancy? Or are they just outcompeted for getting the contract?
> IT projects that cost billions are incomprehensible to me, outside of frontier work
Getting a large bureaucracy to agree on a question where any answer will disadvantage some members of it is frontier work. No-one has a good general solution. Very few people even have approximate solutions for a few special cases.
> is there no space in the market for a good software consultancy? Or are they just outcompeted for getting the contract?
You can't grow big enough to bid on these big projects and stay good.
The people in charge assume that a large scale project requires lots of work to implement in all cases. That is not a given. Creating software that runs 10,000 separate instances is a lot easier than something that has to scale to one instance 10,000 times the size.. There is a bias to 1) assuming the former is what you need and 2) budgeting for the former and 3) preferring the former.
Governments are used to funding a lot of things like transport and the military etc. where costs roughly scale linearly if you are lucky. Usually worse because big projects are so hard to manage and require extra layers of management. If you are really lucky there will be economies of scale.
The people able to bid for big national contracts are also going to be big consultancies who are biased to big and complex projects. For one thing they also have to be able to deploy and support it. maybe if development was separated from deployment and support, it would be better?
There have been multiple examples of things such as smallish government websites (I mean small information sites - the sort of things a freelancer could do) costing millions because its done through a big consultancy.
it's a self-created boondoggle designed to extract money - a scam
that's why the culture created the insane issue of jailing people instead of investigating numbers and fixing bugs - the scam must continue to enrich the players
Don’t forget to mention part of those players are the clients.
“Clients” usually are even worse offenders than these extractive scam companies, mostly because the “clients” are playing with someone else’s money, usually the taxpayer.
The problem for the UK is the Fujitsu owns the last vestiges of the UK IT scene.
ICT/ICL and a bunch of other were all bought out by them in the early 90s.
But also there are two other issues:
1) there aren't many companies that are able to bid on that scale
2) if you remove fujitsu then you're handing it to either crapita or atos. who are also shite.
The solution to this is develop decent in-house architecture team who are able to drive specs and keep an eye on developments.
But
That requires three further things:
1) long term vision, with enough run way/agency to affect change
2) enough money to pay and keep decent people
3) the ability to terminate contracts when they get into shit. This is policitically challenging because contract termination == the current minister is shite in news term. "FLAGSHIP PROGRAMME SCRAPPED BECAUSE OF POOR PERFORMANCE, here is why $government is bad with lots of input from people with agendas that haven't properly been vetted because the journalist isn't a specialist (or its not a news site)"
Part of the answer is dividing programmes up into smaller chunks to attract a wider range of bidders.
The problem is that, even when divided up, they're still often "build and run" contracts which are less attractive to the smaller consultancies which tend to specialise in the "build" part, and are much less keen on the "run" aspect (with good reason, as it tends to be low-margin and painful).
And even then, fashions inevitably change and suddenly someone starts wondering why they've got 30 different tech vendors and need an expensive enterprise architecture team to glue their work together. Cue "why can't we just consolidate all this and get AtoS to do the whole thing?", and the cycle begins again...
Yes, it's all very bad.
> 2) if you remove fujitsu then you're handing it to either crapita or atos. who are also shite.
Yes, agreed, although (at least) Capita are probably much worse. They're currently in a deep hole having screwed over the administration of the Civil Service Pension Scheme. It's been causing deep pain for people for a long time now. Where's IBM when you need them, or Oracle? Just joking.
> Where's IBM when you need them, or Oracle? Just joking.
Palantir? (Also just joking)
Atos has a history with Universal Credit, IIRC.
Admittedly the one time I felt pity for Atos, given that the orders to screw things up came from the client (government)
I find hard to agree they are worse than Fujitsu even with that 'deep pain'.
The Post Office scandal is really really really bad.
A little known but amusing fact:
Fujitsu (well, ICL, the UK company that was bought by Fujitsu and then became the UK branch of Fujitsu) is still the supplier of Microsoft software to the European institutions through a long-standing framework contract with DIGIT.
Why EU could not buy from Microsoft directly to spare the embarrassment is beyond my comprehension.
> Why EU could not buy from Microsoft directly to spare the embarrassment is beyond my comprehension.
Possibly either some form of CYA, or a requirement to buy from an EU company.
Crazy number, crazy software, crazy company, crazy government, crazy scandal, crazy consequences
Why in the world is the ban self-imposed? People literally died because of this, others spent their lifetimes rotting in prison.
A software bug is perfectly fine, everthing has bugs. It was the years of gaslighting, lying and sending people to prison when you could've easily admitted the issue. Ridiculous.
https://archive.is/54nEd
what fraction of the (slow roll) payout is Fujitsu committed to pay by virtue of their collusion in the Post Office scandal?
And Paula Vennells continues to escape real consequences.
And all the rest.
[dead]