That's about $50 million wasted per job.
Also I don't think Oracle should be allowed to hide behind the Phoenix name. This was Oracle PeopleSoft. Why would you pay billions for a piece of software when the maker isn't even willing to put it's name on it?
Pretty sure it was simply a time and materials contract, as is standard. No competent software company will accept a fixed price contract without tightly specified and unchanging requirements, which is something very few customers are able to provide (especially not government agencies).
The issue is it’s hard to scale that up to a multi year 2 billion dollar contract. However, I think individual large scale software projects are inherently a bad idea, which is probably a larger issue.
Just delivering something for the sake of it leads to atrociously bad public infrastructure, be it software or a stretch of pavement.
It is astounding nobody was fired for this.
What would your estimate be to: "rewrite facebook in language L, with the same behavior, bugs and all".
When people think of buying business software systems, they often imagine it's like buying a car or a an app on their phone. A better metaphor for software infrastructure is the renovation or replacement of floors in a skyscraper.
To function properly, business software usually needs to interact or interface with data, software and hardware infrastructure that is in countless existing systems.
The complexity in today's IT infrastructure is mind boggling. It is the result of decades of evolution. The data in existing systems is often held in tens of thousands of individually defined formats and schemas, it flows through thousands of specialized channels that have evolved over years with mechanism for propagation of errors, for dealing with exceptions, for ensuring constraints are in place for security, while allowing special cases when needed. There are rules to ensure redundancy, high availability, backups, to support different user interfaces, different types of computers and mobile devices. There are compatibility layers for different systems, hardware, operating systems, databases. In the case of Canadian government payroll, in the old code there were embedded rules about thousands of different positions negotiated over decades by a bunch of different unions. The business rules in the old code took into account, different programs created in different areas by different governments in different regions.
Skyscrapers can also seem simple from the outside, often being made of flat surfaces forming plain rectangular shapes. But a lot more lies under their skin. There are miles of electrical wires, of plumbing, heating and ventilation, security systems that have to connect and work together. There are stairwells, elevator shafts that have to be planned and positioned for safety.
When parts of skyscrapers get old and outdated, you never see anyone suggesting buying a pre-made piece of skyscraper and shoving it in place of the floors that are outdated. Yet this is what people try to do with software infrastructure.
It's as if an organization having a building with floor 20 to 30 needing updating decided to buy a 10 floor building to put in it's place. They sign the contract and then start working on the details. Putting one building inside another is not easy. You will have to renovate the five floors bellow the new floors and also above, so that the wiring, plumbing stairwell and elevator can connect. The elevators from the two buildings is a particularly difficult problem. They won't connect. As a workaround you might get people to take a detour and a stairwell for one floor to the old elevator, ah and add a two floor high elevator for handicap access next to the stairwell. After a few runs of sewage pipes on the outside wall of the building and running electrical wires in the air crossing from the building next door, your floor twenty to thirty now satisfy the contract and the company that sold you the ten floors considers the job done. Of course it works much worst than the old 10 floors and of course you also now have 20 barely functional floors instead of 10 outdated one. But what are you going to do? You already paid way over budget to hack this mess together. It is now very unstable, if you change anything the whole building might collapse and the company you contracted has pulled a trick. The blueprints to the building (aka. the source code) is proprietary so no other company can work on it.
Undoing the damage would mean completely rebuilding 20 floors (and the building needs to keep operating while this is happening and there is only one company that has access to the blueprints so they can charge anything they want and do a dismal job at it and leave you in a position where you have to perpetually give them more money to keep things minimally functional.
When you go to them because nothing is working, they will steer you towards replacing more of your floors with more of their buildings and break more floor above and bellow while taking away your access to the blueprints and securing their monopolistic position.
Successful businesses never allow themselves to be put in such a bad position for very long. Just recently Amazon announced that they got rid of all their Oracle databases. Governments should start that work as soon as possible. Yes it may require expensive rebuilding of parts of the infrastructure piece by piece, but in the long run it will be much less expensive.
Always renovate room by room, floor by floor while making sure the blueprints remain available so you can use different contractors when one turns out to be a lemon.
I was not proposing the above example as a foolproof plan applicable to every situation, so much as an example of a type of plan that has worked in some scenarios in the past.
I should also point out that part of the plan for the Scandinavian plan wasn't the replacement of government workers with locals, as much as the expectation that by moving so many jobs into such small areas, they were guaranteed to majorly shift the local economies. Numerous jobs opened up overnight supporting all these well paid government employees moving into the area.
Nonetheless, I am happy to learn about the counter-example, thank you!
It would be nice if you could name the failure you cite, too.
Government experts will live in a lot of places. NASA has major operations in a metro area in Alabama with only ~500k people, for instance: https://en.wikipedia.org/wiki/Marshall_Space_Flight_Center. That's more periphery than center, in my mind.
Sorry, should of course named it as that's what I was asking. It was the move of the Finnish Medicines Agency from Helsinki to Kuopio. Checking the latest status, they do now have presence in both cities so it wasn't cancelled in the end; I guess they pushed on but ended keeping the Helsinki office for the experts refusing to relocate.
Do you have the name of your example?
Except: 1) UBI is unproven and 2) it strikes me as a bit of a band-aid solution designed mainly to avoid confronting certain ideological assumptions.
1. doesn’t cover the rising costs of medical care and insurance (in the us at least)
2. won’t keep up with inflation
3. sets up private markets to potentially inflate prices “because people have 1000 dollars extra to spend”
UBI in concert with price limits for basic goods and services, insurance reform, and medical costs limits might have a chance.... but then you are now confronted with the potential that it’s now the “rich and large corporations” who are subsidizing everyone, giving them extraordinary political power over the govt/people
i’m still skeptical of UBI
1. UBI (including some fractional amount per child) 2. Free universal healthcare (i.e. Medicare is no longer age-limited) 3. Better education & free public universities 4. Some kind of higher tiers of UBI for disabled people who really need it (judged strictly)
And then get rid of every single other means-tested benefit (including Social Security), as well as tax-advantaged retirement plans (which overwhelmingly benefit the already well-off), and get rid of the favorable tax on LTCGs and carried interest; everything pays full-freight income tax. After this, benefits are either free to everyone (i.e. healthcare) or you pay full freight out of your UBI. And UBI is not scaled by CoL, because it doesn't make sense to subsidize landlords in expensive cities.
And even with all of that, it's not clear that it won't be vastly more expensive than the current system, making it effectively unaffordable. There's also the issue that some people will not spend their money correctly, e.g. they'll fritter it away on drugs/alcohol and not have enough to pay rent or buy food. So you'd need some kind of combination of food stamps and Section 8 housing vouchers, paid out of UBI, for people who need to be forced to allocate their money correctly (as there will no longer be any separate means-tested food stamps or housing assistance).
I could see something like this working in smaller, well-connected, yet independent communities, but I'm struggling to see how it could scale up to a country the size of the US. The rich would be paying so much more in taxes to essentially subisidize many people who might choose not to work at all, that I think it's inevitable they'd manage to overthrow it.
Like you, I think from a realpolitik view of things (if nothing else), that it can't work.
How much would UBI reduce the crime rate? How much would it increase the rate of entrepreneurship? How much would it reduce anxiety and depression? How much would it increase average education levels? Upward mobility? Democratic participation and voter turnout? Volunteer work and charitable giving?
No one knows the answers to these and many other similar questions, but it seems plausible to me that the cumulative effect could produce drastically more wealth than it costs to implement.
Our country (not to mention our world) is facing many extremely severe and worsening problems, many of which are rooted in economic inequality. We are, in my opinion, highly unlikely to solve these problems by incrementally chipping away at them, and the prospects for leaving them unsolved look increasingly bleak. So I think we need to be willing to consider big solutions if the cost-benefit analysis is favorable, even if they do require huge investments. We know UBI is technically affordable based on the amount of wealth that exists in our country—the question is whether it would really work (something we can test), and then if it would, whether we can summon the political resolve to make it happen.
Not attempting such things doesn't mean we avoid the risk of such wealth destruction; it just means we have nobody to a point a finger at when it happens. Politically this is easier because of how people think and the dynamics of our social discourse. See Appeal to Nature Fallacy (https://en.wikipedia.org/wiki/Appeal_to_nature). This is especially true the more cynical and ideological we become as we're more likely to pounce on failures of action, but excuse failures of inaction as inevitable natural events or limitations.
> Such a strategy failed utterly in Canada.
It's possible the strategy is good, they just utterly failed at executing it for other reasons.
A challenge is that many solutions to problems like unemployment have difficult to quantify primary effects - e.g. actual jobs created - while they have easy to quantify opportunity costs - expected return of investing capital in ETFs, or something.
My point is not the details, or to say that creative solutions can’t return more to the community or country than the financial costs of creative solutions - merely that those in the positions to propose these solutions are frequently unable to adequately model the return.
In another post a few months ago, there was a discussion that the government shouldn’t have to propose an ROI (in the context of building new infrastructure)- but it was pointed out that frequently these projects are financed with debt, and that lenders dictate a return.
If we want to move away from ROI calculations, and worrying about unquantifiable primary effects, a critical component IMO is that we can’t really pay for them with debt.
If you pay with debt, you are beholden to a real, quantifiable cost of capital. If you fund with cash, or equity, it gives you the freedom to pursue preference over return, or theory ahead of evidence. Startup funding works this way for a reason. Creative solutions require freedoms not always afforded by a time-bounded and certain cost of capital.
ROI is just one kind of utility function, but converting everything to said resource-equivalent works pretty well.
It's always a nasty burden to decide between hospitals and roads, but it's a neccesity also, because building just hospitals without roads doesn't help much, because ambulances can't get to them. And vice versa.
Sure, ROI in itself is a bad metric, and the problem is that developed nations are frequently get caught in an overly simplistic financial modelling situation. Because politically a more complex model is very hard to defend and the simple model gives simple answers, which are usually not optimal.
As you said, creative solutions are what's needed, but they are harder to quantify.
But there's always the risk of a pathological co-dependency on government employment. You can see this in many developing countries. So for partly legitimate reasons, the Civil Service has long been the boogeyman of Republican politics. Specifically, that civil servants are reliable "big government", Democratic voters. Better to 1) pack their voting power and 2) prefer private out-sourcing. This is one major reason (if not the major reason) why government agency activity and policy have migrated to D.C., and why the headquarters of so many corporations have moved to the D.C. metro area.
Those that choose not to move can leverage their experience in private industry.
My point is that creative solutions are possible, if we dispense with the notion that the only solution possible is the free market.