Slashdot Mirror
New NASA Launch Control Software Late, Millions Over Budget (go.com)
schwit1 writes: The launch control software NASA is writing from scratch for its Space Launch System (SLS) rocket is way behind schedule and way over budget. According to ABC News, "Development of this new launch control software is now projected to exceed $207 million, 77 percent above 2012 projections. The software won't be ready until fall 2017, instead of this summer as planned, and important capabilities like automatic failure detection, are being deferred, the audit noted. The system is vital, needed to control pumps, motors, valves and other ground equipment during countdowns and launches, and to monitor data before and during liftoff. NASA decided to write its own computer code to "glue together" existing software products a decade ago -- while space shuttles still were flying and commercial shippers had yet to service the space station. Both delivery companies, SpaceX and Orbital ATK, rely on commercial software, the audit noted."
In other words, even though NASA could have simply purchased already available software that other launch companies were using successfully, the agency decided to write its own. And that decision really didn't come before the arrival of these commercial companies, because when it was made a decade ago that was exactly the time that SpaceX was beginning to build its rocket. This is simply more proof that SLS is nothing more than a pork-laden waste of money designed not to explore space but to generate non-productive jobs in congressional districts.
In other words, even though NASA could have simply purchased already available software that other launch companies were using successfully, the agency decided to write its own. And that decision really didn't come before the arrival of these commercial companies, because when it was made a decade ago that was exactly the time that SpaceX was beginning to build its rocket. This is simply more proof that SLS is nothing more than a pork-laden waste of money designed not to explore space but to generate non-productive jobs in congressional districts.
That last few sentences were really inacceptable. Could someone edit this?
There is no software on the planet that is more scrutinised and more meticulously developed than software for spacecraft.
Start a Softwareproject like that without having it properly planned or the right people involved and your project will go over budget manifold inmediately.
No surprise here.
We suffer more in our imagination than in reality. - Seneca
What kind of summary/ story is this? This was only the executive summary. We want to know what systems the stuff runs on, what programming languages are used; you know, the geek stuff!
even though NASA could have simply purchased already available software that other launch companies were using successfully, the agency decided to write its own.
NASA decided to use existing software that was known to work and was fully understood rather than rely on commercial software which could be total shit. besides, they wouldn't be purchasing the software, they would be licensing it which means they would likely have to pay $X for Y computers for Z years. also, what happens when they want to add feature XYZ and they are unable to? freedom isn't free... it's 207 million dollars.
Anons need not reply. Questions end with a question mark.
. . . .and when the FBI started to develop its' own case manager, the "Virtual Case File", which was one of the more spectacular failures in Government IT Development.
When the post-mortem finally comes in, I'd be more than willing to bet that it was due to (1) lack of formalized baseline requirements to hang an initial design on, and the real program-killer, (2) constant requirements creep. Because contractors are unwilling to tell a Federal Customer "no" (because it usually results in decreases in funding in the next task order, or re-allocation of slots to another contractor. . .), there's a constant "just add this one little thing". Over and over again, until you have an unworkable mess and a design that looks nothing like the initial requirement.
The same kind of pressures destroyed the Navy A-12 "Avenger" attack jet in 1991: constant scope creep, until the aircraft was too heavy to fly off an aircraft carrier. The resulting legal fight lasted 13 years. . .
28-page PDF
Yes, I'm sure you're going hungry because an average of $0.20 per year over ten years was sent on learning about a whole new class of planet that may represent the dominant form of equilibrium celestial body in the universe.
Things like NH are exactly what NASA should be doing.
Don't get me wrong, I strongly oppose SLS. But you know, it's perfectly understandable that they keep ending up going down these roads. They have too much infrastructure and too much personnel focused on building large rockets. The infrastructure especially, as you can't just "lay off" (or phase out) it like you do with people, and it costs you money to maintain. And the people making the decisions - congressional representatives - aren't experts, they're just ordinary people, and thus easily swayed by arguments made by advocates of these jobs (direct and indirect) in their districts.
It's understandable. But it needs to be fixed.
There's long been resistance to the privatization of rocketry. We remember the first privatization battles back in the Shuttle era, and how much resistance there was to letting Atlas and Delta go private rather than just disappear altogether. Indeed, the new battle in much of the launch market isn't so much over privatization as it is over "old school" vs. "new school" private companies, with the former offering evolved expendable launch vehicles with generally good track records but high price tags, and the latter offering ground-up vehicles with short launch histories but very low price tags.
Regardless of how this battle goes, there really isn't a place for NASA in it, any more than there would be a place for a government car maker to compete with giants like GM, Ford, etc and upstarts like Tesla in terms of making passenger cars. There certainly could be grounds for a government agency to conduct basic research that can advance automotive technology, things that private companies wouldn't pay for because it doesn't provide a short-term fiscal payoff or would help their competitors as much as it would help them - public health and safety, advanced concepts for the future, etc. But they shouldn't be making cars.
The same applies to NASA, only moreso. There's a *lot* of basic research in the aerospace industry that's either too expensive for its risk level or only associated with long-term payoff for them to conduct. And this is where NASA should be: planetary science and advanced concepts. But getting there is difficult, as it means having both an administration and congress who recognize the need to reorient NASA and are willing to accept the economic pain involved in doing so. It doesn't necessarily mean budget cuts - but it means closing facilities, selling off hardware for well less than you paid for it, and jobs leaving certain areas, even while new facilities open, new people are hired elsewhere, etc.
SpaceX wants to take over the private launch industry and use the money to go to Mars? The reaction shouldn't be "Oh noes! That's our goal, back off!", it should be "Awesome, that will save us a ton of money! How can we help?" And then let them spend their money on the glamorous stuff while NASA works on the less glamorous stuff behind the scenes. "That's a nice looking gigantic methane-fueled engine you've got there - we could hear it for miles when it fired up! So anyway, what do you think of this long-lifespan dirt scoop we made to dig up mucky ice for the habitat? And this water nanofiltration system to maintain electrolyte concentrations in the necessary levels in the electrolysis unit - want to see it? Oh, you're too busy playing with engines? Oh okay..."
Hourglass says she knows a kid in Iowa who grows up to be president.
The "outsource everything" mantra of the 80s is still with us. It doesn't say in the article, but most of the work done by NASA is actually performed by outside contractors, and wI would bet a dollar that this is the work of USA (United Space Alliance - aka Lockheed and Martin Marietta) or one of the other giant government contractors like CSC or Booz. It may be NASAs project, but congress pretty much has gutted the real workforce so everything merely has project managers rather than actual engineers.
As for the submitter's (and, to some extent, the article writer's) take, I think they got it backwards. This project was started BEFORE there was any commercial launches of significance and so the code simply didn't exist for a robust launch control system as envisioned. The article does point out that there is more software available today, and that it could be an option. OTOH, we're talking about proprietary code from one of two competing firms with no outside review of the codebase. That's fine for putting up a couple of tons of food and electronics, but a private company has yet to successfully, reliably put humans into earth orbit. And that kind of responsibility is an order of magnitude higher than supplies.
Is it just my observation, or are there way too many stupid people in the world?
using the same argument, we should certainly be privatizing the military.
A couple of key points: 1) Software development at NASA is unlike anything 99.999% of the tech heads here would recognize. The scrutiny and the level of detail and failover protection in their code is unbelievable. Lives are at stake with their code or at worst millions of dollars of hardware crash into the surface of Mars and an entire mission is ruined. For most of us, if we have a bug in our code we patch and life moves on. If NASA has a bug in their code someone may die. 2) I have a friend that works for NASA. After SpaceX's first successful launch I said to him "big deal, NASA gave them blueprints, designs so this wasn't much of an accomplishment". His reply was "no, NASA did go to SpaceX early on and offered assistance in design but SpaceX turned them away. SpaceX is in business to make money and they cannot carry the burden of triple and quadruple redundancies that NASA has in their spacecraft. NASA's designs are too expensive." So if SpaceX is using off the shelf software you should recognize that SpaceX is willing to accept defects in order to ensure their profitability. NASA's needs are fundamentally different; they are to protect life and missions. SpaceX will take risks that NASA will not(and BTW, go ahead and sign up for that ride on a SpaceX rocket knowing this). This is why I am sure NASA did not choose off the shelf software and decided to write it themselves. This post is way off the mark to call this just a pork play.
Ahem....no. That's the per capita cost. Total cost for the mission is about $720 million, or about $2 per person.
Comparatively, the F-35 Joint Strike Fighter program cost is $1.3 trillion, the Boston Big Dig cost $24 billion ($7 billion was federal money), the new Vikings stadium is expect to cost just over $1 billion (and will probably run way over). New Horizons was a bargain and exactly what NASA should be doing, not SLS.
The more people I meet, the better I like my dog.
> It's a frequent refrain that government agencies are overspending and missing deadlines.
The implication being that private companies don't ever have budget and schedule problems. Which is -- of course -- quite untrue. They have the same problems as the government. All the time. Read up on the Trans Alaska oil pipeline some time.
I'd suggest the big projects will always tend to have problems. The reason is simple. Scheduling is done using the most probable time for each task. But some tasks take more time than projected and some take less. The distribution is skewed. Sometimes a "one month" task only takes two weeks. But perhaps equally often it takes two months. String together a dozen or so complex tasks and you can bet on usually taking longer than you expect and spending more than you budgeted.
And that's when you do the planning well ...
You can't see ANYTHING from a car, You've got to get out of the goddamned contraption and walk...Edward Abbey
...even though NASA could have simply purchased already available software that other launch companies were using successfully, the agency decided to write its own."
Yeah, other, incorrect words. "...when it was made a decade ago that was exactly the time that SpaceX was beginning to build its rocket." Because when a commercial rocket is being built is the time to decide it's software is ironclad? And that the software you have available from NASA that was tried and true was definitely defunct? NASA has a lot of process to go through, including COTS and acquisition, so I'm not sure what you're proposing here was a better alternative without hindsight.
Yes, the article seems to be trying to manufacture outrage out of nothing.
It's a make-versus-buy decision. Industry does these decisions all the time. When your applications are unique, the decision tends to go toward "make your own;" when your application is something that many other people also do, the decision tends to go toward "buy the commercial product".
Buying off the shelf comes with hidden costs unless what is available exactly meets your need-- if you need to write a new contract for every change (and since you still haven't designed the system you're launching, there will be a lot of changes needed, as you keep refining requirements) every single change is a chance for the vendor to demand large dollar payments.
And the article's statement "why doesn't NASA just use what Space-X used" is absurd. Ten years ago, Space-X was an unknown company who had just launched their first rocket. Which failed. As did their next launch. And the one after that.
http://www.geoffreylandis.com
We did. That is why we had issues in a number of these nations
I prefer the "u" in honour as it seems to be missing these days.
No, the submitter has 3rd grade reading comprehension skills and an agenda. And it's in the summary which suggests that software was available, TFA quote is actually "Commercial software products would be a better option for NASA as well, according to the audit, especially given recent advances in the area." Note the use of future tense, not past unreal conditional. IOW, it still might be more economical at this point, but it wasn't an option when the project was started.
The commercial firms which are servicing the space program had never delivered a payload to orbit at the time of the original design specification and plan, and had no tested software at the time. NASA, otoh, had subroutines already written which had been tested and vetted for decades, over hundreds, if not thousands, of successful launches and NASA wanted to use those [tested] routines in their new launch control system. The contract was to put them all together into a cohesive whole. Not a trivial task, but also not writing from scratch.
More importantly there is no COTS software in this arena. This is not Word with a custom skin. This is piece by piece built based on the unique hardware and control systems which are part of the critical safety path of the launch sequence. Even if SpaceX or another space transport company has software they use, it would have to be stripped, rebuilt, and re-tested for the configuration at the KSC launch complex to be used for these flights.
Is it just my observation, or are there way too many stupid people in the world?
> There certainly could be grounds for a government agency to conduct basic research...
I would say basic research, AND tactical purchasing contracts to motivate hardware manufacturers to develop to broadly useful standards and capabilities.
For example, they could order the first batch of "orbital tugboats" designed for permanent orbital service, including maintenance and refueling. Vehicles which could be broadly useful platforms for a wide variety of services, but would involve some very different design compromises than existing vehicles. Once NASA has absorbed the initial development costs, the same basic vehicle can then be sold to everyone else at much closer to production cost. Include in the specifications a need for specific royalty-free docking hattches and external mounting points and you've also just created a de-facto standard for all other manufacturers to build to.
--- Most topics have many sides worth arguing, allow me to take one opposite you.
Indeed. I'm a fan of COTS-style programs as well. NASA can ensure a market for something in the beginning while giving the actual market time to develop.
What I wrote isn't to mean that NASA should "get out of rocketry" entirely. The key issue is, if they're doing rocketry, it should be to innovate - as mentioned, advanced concepts. To develop fully functional technology demonstrators rather than workhorse deliverables. To build on as small of a scale as possible for a given new technology, with full acceptance that project failure might occur. Airbreathers, metastable fuels, ballistic launch, active suspended launch structures, composite cryogenic tanks, inflatable reentry, you name it... any untested technology that claims significant benefits but is too risky/expensive for private enterprise to develop from scratch, that should be NASA's domain. But it shouldn't be approached with the intent to make a workhorse. We wouldn't accept the NHTSA making and selling cars and trucks, the FAA making and selling passenger jets, or the DOE making commercial power plants - and nor should we accept NASA doing the same with rockets.
Hourglass says she knows a kid in Iowa who grows up to be president.
To be fair, Alan Stern called the four years he spent leading the development of New Horizons a blur. That was half the length of the typical development cycle for comparable probe/satellite projects, and put an enormous personal and professional strain on all the people working on it. It is not reasonable to expect people to perpetually work at that pace. The New Horizons program was a bargain largely because it put undue and unsustainable strain on the employees and contractors involved. Yes, it was a well-managed program and is a good example of cutting a lot of red tape to get a project done, but you should expect a comparable program to cost half again that much if you're not overworking your employees. The only reason the project had such a short timeline in the first place, IIRC, was because the calculated launch windows were so rare.