After 16 Years and $8 Billion, the Military's New GPS Software Still Doesn't Work (arstechnica.com)
- Reference: 0181186718
- News link: https://tech.slashdot.org/story/26/03/30/1950214/after-16-years-and-8-billion-the-militarys-new-gps-software-still-doesnt-work
- Source link: https://arstechnica.com/space/2026/03/after-16-years-and-8-billion-the-militarys-new-gps-software-still-doesnt-work/
> Last year, just before the Fourth of July holiday, the US Space Force officially took ownership of a new operating system for the GPS navigation network, raising hopes that one of the military's most troubled space programs might finally bear fruit. The GPS Next-Generation Operational Control System, or OCX, is designed for command and control of the military's constellation of more than 30 GPS satellites. It consists of software to handle new signals and jam-resistant capabilities of the latest generation of GPS satellites, GPS III, which started launching in 2018. The ground segment also includes two master control stations and upgrades to ground monitoring stations around the world, among other hardware elements.
>
> RTX Corporation, formerly known as Raytheon, won a Pentagon contract in 2010 to develop and deliver the control system. The program was supposed to be complete in 2016 at a cost of $3.7 billion. Today, the official cost for the ground system for the GPS III satellites stands at $7.6 billion. RTX is developing an OCX augmentation projected to cost more than $400 million to support a new series of GPS IIIF satellites set to begin launching next year, bringing the total effort to $8 billion.
>
> Although RTX delivered OCX to the Space Force last July, the ground segment remains nonoperational. Nine months later, the Pentagon [1]may soon call it quits on the program . Thomas Ainsworth, assistant secretary of the Air Force for space acquisition and integration, [2]told Congress last week that OCX is still struggling.
The GAO found the OCX program was undermined by "poor acquisition decisions and a slow recognition of development problems." By 2016, it had blown past cost and schedule targets badly enough to trigger a Pentagon review for possible cancellation.
Officials also pointed to cybersecurity software issues, a "persistently high software development defect rate," the government's lack of software expertise, and Raytheon's "poor systems engineering" practices. Even after the military restructured the program, it kept running into delays and overruns, with Ainsworth telling lawmakers, "It's a very stressing program" and adding, "We are still considering how to ensure we move forward."
[1] https://arstechnica.com/space/2026/03/after-16-years-and-8-billion-the-militarys-new-gps-software-still-doesnt-work/
[2] https://www.airandspaceforces.com/pentagon-eyes-canceling-troubled-gps-ground-system/
Isn't that what AI is for? (Score:2, Funny)
Why don't they just ask Claude to fix it?
Better yet, give Elon Musk a couple hundred billion to start a private company and hire all the people, then hire that company to complete the project. Then it will be ready "this year".
Vibe code it! (Score:2, Funny)
I'm sure AI agents you deploy from your meta glasses can fix that in a weekend.
Aerospace FFRDC role? (Score:5, Interesting)
Aerospace (and other FFRDCs like MITRE) exist to prevent massive failures like this. I wonder what the Aerospace corporate explanation is. I know from working at another FFRDC that often the worker-bees know the program is heading to disaster, but the managers won't carry the bad news to the customer. Other times, the bad news is delivered, but the government manager decides to carry on anyway. That can be due to pressure within the Service ("Don't f**k this up!, Colonel!"), or pressure from the contractor ("Trust us, these problems are temporary.")
The causes are often requirements instability, overly ambitious/unimplementable/unrealistic requirements, impossible initial schedule ("1 month to make the baby with 9 women"), technology problems (immature technology, vendors can't deliver as promised), and occasionally manufacturing/assembly/integration problems. And of course, substantial amounts of the functionality is in software, and this community knows the ways software projects can go south.
It's no consolation to this project,, its leadership, prime contractors, and customer community, but the last major project I worked on failed at 2 1/2 times the sunk cost of this one.
Re: (Score:2)
and TFA is a "there at it again..." type rage piece without giving even a taste of what many things really went wrong along the way from a design or other tech POV.
Re: (Score:1)
> and TFA is a "there at it again..." type rage piece without giving even a taste of what many things really went wrong along the way from a design or other tech POV.
so? this should never happen.
whatever the reasons this is Category 5 Grade A incompetence.
the goddam Pentagon has FAILED 8 audits in a row.
i guess we can blame all that on DEI and everything will be rosy now that the Whites are back in charge.
Re:Aerospace FFRDC role? (Score:4, Informative)
Failing audits is frankly independent from failing programs. The audits usually have problems tracking money flows and then property within the government. The contractor's expenditures are closely monitored. That doesn't mean they're in-line, but they're auditable. And when the audit discovers problems, there are ways for the government to respond. I've seen those applied rather frequently.
One common pattern is a program starts down the wrong path, and blows initial cost/schedule/performance. But that capability is needed badly (often because its predecessor program didn't deliver). So the Service piles on more requirements and 'readjusts the baseline' for additional funding, because "if we don't get it in this Program of Record, it'll be at least a decade before we can start a new Program of Record to get what we need." That just adds requirements to something that is already behind. If I had to guess what happened here, I bet there's some of that flavor over the execution. In my experience, most programs started with the combination of unachievable or under-specified requirements AND unachievable/unrealistic schedule.
(A 'Program of Record', by the way, consists of an approved requirements document, an approved POM budget for the next 5 years showing the RDTE money, the OPA purchasing money, and the OMA maintenance money FOR EACH FISCAL YEAR. If you run out of RDTE money but haven't finished the design, you're in trouble. The third element is the approved procurement strategy, that says how you'll buy it. That includes the kind of contract, firm fixed price or cost plus, the kinds of oversight, when and how prototypes will be delivered and tested, etc.)
Re: (Score:2)
>> and TFA is a "there at it again..." type rage piece without giving even a taste of what many things really went wrong along the way from a design or other tech POV.
> so? this should never happen.
> whatever the reasons this is Category 5 Grade A incompetence.
> the goddam Pentagon has FAILED 8 audits in a row.
> i guess we can blame all that on DEI and everything will be rosy now that the Whites are back in charge.
What part of "this" should never happen? I'm not a fan of massive contracts going over time and over budget... and this particular one seems to be an outlier, but the reality is that the majority of these large acquisition projects go over time and over budget and some of them outright fail. And it's been like this at least since "The Mythical Man Month" ( [1]https://en.wikipedia.org/wiki/... [wikipedia.org] ) was first published 50+ years ago.
So just saying, as TFA does, "look at this project that's like 3x over duration
[1] https://en.wikipedia.org/wiki/The_Mythical_Man-Month
Re: (Score:2)
If it's a cost-plus contact, the incentive is to milk it for as long as possible, without delivering a product. It's the definition of a perverse incentive.
It's called corruption (Score:5, Insightful)
Eisenhower knew the deal in 1961.
Re: (Score:2)
Aside from the general myth of privatization being more efficient its misaligned incentives - the company gets paid more the longer it takes, they aren't incentivized to hire (really outsource to) competent programmers.
I wish it was corruption - it's bad management (Score:5, Interesting)
You can fix corruption. I've been an embedded programmer on constrained devices for over 30 years. Mostly as a contractor fixing messed up projects. The quality of embedded projects slowly declined in the 2000s and 2010s but has fallen off a cliff over the last 5 years. It used that the quality of embedded developers was higher than enterprise developers (there was a higher minimum bar just to get your code to run). It used to be that you had to plan and organize your projects and the projects were run by engineers. You used to get support from your suppliers. In the 2010s part suppliers gave up on support and direct you to forums. When you do get documentation from suppliers you have to guess at which parts are correct and what steps they have left out. I had one formerly reputable supplier fail for 3 months to get "Hello World" to work on their dev board.
Then there is the management of projects. Now the de facto person in charge is the one in charge of the Jira tickets. They decide what tasks get resources and they decide which engineers do what tasks. They can't understand a development plan, they can't build the code, they usually can't even use the product. What they can do is take bugs from testing, create tickets and assign engineers to the tickets. Their only metric is how fast tickets are resolved. They don't care if the engineer they assign knows nothing about the project
The most successful embedded engineers today
Don't put useful comments their code (they may describe how the code works but the code says what it does)
They never document (and most projects don't even have a single place to put documents or have a way to find them)
They are good at volunteering for easy or high visible tickets
They close bugs by creating global variables that track the condition of the bug, then adding a function that is called all the time, a function that then checks the globals for the error condition, prevents (or masks) the error and then (hopefully) cleans up the globals without creating too many other bugs.
- This means they never have to understand the code. If there is automated testing they might never even need to know how to use the product.
Re: (Score:2)
Is the rise of cheap but powerful embedded processors and the Arduino ecosystem partly to blame? I don't think I've seen very much high-quality and efficient arduino programming done. C++ with lots of unnecessary OOP. As long as there's a watchdog timer to reboot the thing every couple of hours, she'll be right. Ship it. Heck I use micropython for a lot of projects lately. It's very cool but it may not be preparing young developers for the rigors of aerospace development.
Recommended reading (Score:5, Insightful)
"The Spoils of War" by Andrew Cockburn. Goes way back to American soldiers having to steal boots off dead Chinese soldiers in Korea to get decent boots, their feet were freezing off.
"The Pentagon Wars" by Col. James F Burton. Burton was part of the 1980s "Fighter Mafia" who got the F-16 built, against Pentagon tendencies for every new plane to be twice the weight and twice the cost of the last one. (The F-35 continues the tradition.) They were the ones who publicized the $400 hammer and $600 toilet seat.
Cockburn laments that people thought it only applied to some things, when their point was that every $1 lightbulb on the console was $25 to replace.
Burton notes that one Army logistics guy got the price of a single uranium bullet down from $80 to $4 by whipsawing two suppliers into real competition, another reduction every purchasing round, for years. That guy gave a presentation to a roomful of Stars on it, and came back to his desk to find retirement papers waiting. Or a transfer to Thule. His choice.
Re: (Score:2)
Those "Stars" that fired him need to have their arses handed to them. Corruption is something I fear we will never be able to solve, but I live in eternal hope.
This How To Fix The Problem (Score:4, Funny)
Turn the project over to the Ukrainian they'd get it fixed so they could use the full military precision version of our GPS for themselves.
Re: (Score:3)
> Turn the project over to the Ukrainian they'd get it fixed so they could use the full military precision version of our GPS for themselves.
This is likely way to true and they'd do for 5% of the cost.
Re: (Score:1)
I would like to say how cool your Google+ login is. It shows up in your post like a bright red beacon. I once even wrote a haiku-poem for Google+ login icons on Slashdot. Take it! It is yours!
Google Plus Login
red on a green Slashdot sea
setting my soul free
exotic matter
how I long to fly with you
Google Plus Login
Google Plus Login
fancy Google Plus Login
Google Plus Login
primitive tribesmen
gaze at the little red square
dream of things to come
I could do it cheaper (Score:4, Funny)
I'm not a programmer, but I could make a system that doesn't work for only $4B. I'm only discounting it because I want to save taxpayer money.
Where is DOGE now? (Score:2)
Subject is the question. Where is DOGE on the big stuff? The Pentagon wastes more every month in fraud, waster, and abuse than USAID spends annually. But somehow charity gets the axe and Ratheon keeps getting multi-billion dollar contracts, no strings attached. Can anyone put aside the woke distraction and look at the serious problems?!?
Re: (Score:2)
100% We need anti-fraud to serve the taxpayers, rather than just pander to political "culture war" objectives. But to be fair that was valid too - no one can grift like virtue signallers. But yeah - there is a shitload of other areas to investigate - especially the military industrial complex.
The cart before the horse (Score:2)
The fact that they've launched more than 30 satellites without working software is a mind-boggling example of poor planning and non-existent oversight. I know you need more than one up there to test the software, but if they couldn't get a fully working proof of concept covering just CONUS with ten then they should have stopped paying for launches.
Re: (Score:2)
They can update the software over the air. But launching something that has a limited orbital life ahead of time is pretty stupid.
Wrong programming team (Score:2)
They should have hired Ukrainian programmers.
Behind schedule, over budget (Score:2)
In federal contracts, it is not unusual that a project is behind schedule and over budget (sometimes it is not actually the contractors fault, for the project scope changes during the period of the contract). But this is quite the example of why the government needs to contract smarter.
Oh but it works very well (Score:4, Insightful)
It works wery well at it's intended purpose of griefing taxpayer money.
Re: (Score:2)
If only I had mod points!