NASA Will Have To Wait For Mars

mattg writes, "Auntie is covering NASA's timetable for recent explorations of Mars has been called "wildly optimistic". Dr. Carl Pilcher, leader of NASA's planetary exploration program (whose sweater at the time said "Obey gravity: it's the law") has admitted that they do not know if they have the technology to bring rocks back yet. The report into the loss of the Polar Lander is due out at the end of the month. "

Comments

Re:I can see it coming

[bughunter]

It gets worse than that. I've worked as a systems engineer for 10 years on NASA contracts, and lately, the "faster, better, cheaper" has exacerbated what I have always considered to be the worst problem in the aerospace industry: underbidding.

When NASA issues a request for proposal (RFP), the bidders have a good idea of what the proposed cost should be in order to have a competitive proposal.

In the old days, programs were "cost plus fixed fee" (CPFF). In other words, whatever the cost of the project in the end, the customer (NASA) would pick up the cost, and the contractor would get an additional fee on top of that (gotta make a profit, of course). But there were a lot of abuses of CPFF proposals, so there are few left (mostly DOE nowadays - check out the Savannah River operating contract). I never had the leisure of working on such a program, but I have heard some "war stories" from the older engineers, and some of the abuses were astonishing.

So nowadays, programs are fixed cost. The original idea was to force the contractor to agree to a fixed payment for the program, and that price would have to include any profit that the contractor hoped to make. That lead to problems not with overbidding, as one might think, but to "no bids" and failed contracts due to cost overruns. So it was tweaked and the current policy is a fixed price contract, plus performance awards based on the programmatic, technical, and financial performance of the contractor. The cost of performing the work is agreed upon, and then NASA establishes another amount as a "carrot" to induce the contractor to perform well. If NASA doesn't like the contract performance, they can withhold part (or all) of the carrot.

It works pretty well for NASA, so far, so they haven't changed it in the past 6 years or so... but on the contractor end, it leads to two things: underbidding on contracts to insure some profit, and overworking the engineers to maintain performance.

The underbidding almost always comes in the labor category. In the task estimation process of the proposal, one "chunks" the project into small tasks like "design dunselhickey firmware," "design dunselhickey electronics," "design dunselhickey mechanical and packaging," "integrate and test dunselhickey," where the dunselhickey is an attitude control subsystem, or a sensor instrument, or something. (And I'm ignoring the contractor/subcontractor/vendor hierarchy to keep this somewhat short.) For even the simpler systems like Deep Space 2, these task estimates are huge efforts, and whole forests are sacrificed to them. Anyway, the point is that the contractor management knows ahead of time how much they want to quote for cost, so if the estimators (the engineers) don't come up with a small enough number, the managers (accountants, lawyers, and engineers with 30-year-old training) take a chainsaw to the estimate to trim it down to their target cost. When it comes time to perform the contract, the engineers find that there's not anywhere near enough money budgeted to perform the labor that needs to be done.

Which leads to the next problem: overworked engineers. The contractor who wins the project faces a dilemma as work begins to fall behind schedule. Contingency was never a part of the budget, so any delays or technical problems, even in the early phases, directly impact the bottom line/delivery date. And in almost every contract, there are several areas where the budgeted money to perform the work is grossly inadequate. In order to avoid cost overruns and keep their performance award, management puts more and more demand on the engineers to take shortcuts and work overtime. Unpaid overtime, of course. Which leads to fatigue and the resulting errors and oversights, as tesserae described. And of course, they're always the engineers' fault. (As I like to say, "parts are derated; engineers are berated.")

Faster, Better, Cheaper has only made this problem worse. There's less money budgeted for any given doowidget, but more performance demands. The leadership is out of touch with the technical demands of the performance requirements, and promise more for less. The technology only does what we tell it to do; if we take shortcuts in design and testing, we don't know what we're telling it to do. Engineers want to do things right, and know they can do things right the first time, but the available time (e.g. money) has been shrinking steadily.

But at times like this, when I'm feeling most cynical, I can still take solace in the fact that I'm not working in a competitive commercial environment (like application software) where the situation is even worse. When I see that Win2000 shipped with 64,000 "issues," I know exactly what's going on... the politics and jargon may be a bit different, but it's still management's fault for promising more than they can deliver.

MTV's the "Real World-Mission to Mars"

[yuriwho]

At this rate an entertainment company will be the first to get a manned mission to Mars. The mission will be paid for by a 24 hour, cable/satelite channel that broadcasts the entire mission complete with space sex (pay per view for that tho) after the audience has developed close personal relationships with each of the characters on the mission (a bunch of photogenic 20 something astronauts) we will all get to watch them crash into Mars..live..in the greatest rating event ever. Given the extreme financial success of this mission, the sequel show will be launched immediately, this one lands and then everyone starves to death in a gripping drama lasting months with a strange plot reminiscent of Lord of the Flies.

At least we would have landed humans on Mars.

I can see it coming

[tesserae]

The thing that annoys me the most about the loss of the last two missions -- and the thing that is probably not going to be blamed for the losses, in the end -- is that the hardware was actually pretty much up to the job; it's just the handling of the hardware by the humans involved that cost us those missions, and possibly this whole exploration campaign.

The Climate Orbiter was lost because two people (one NASA, one from the contractor) were handling the entire trajectory; they were completely overworked (to the point failing to implement the backup planning which was already on the timeline, and which by itself might have saved the mission), with no one to even do basic sanity checks on their work -- and they missed not only the critical units conversion, but also the fact that their trajectory corrections weren't having the desired results. A college kid on a work-study internship, working ten hours a week, could have saved the mission. But it was faster-better- cheaper , so they didn't hire the kid...

The Polar Lander appears to have been lost over communications failure between two test groups: when the lander legs were dropped, they apparently rebounded and triggered a ground-contact sensor in each leg; this set a bit in the computer, so that it "thought" the vehicle had already touched the ground, and it killed the engine as soon as it took control. The rebound happened regularly during testing, but the group testing the leg deployment didn't look at the bit's value at the end of the test (after all, it wasn't on the ground yet, so it wasn't their job...); and the group testing the final powered descent didn't bother to look at the contents of the register before they started the test -- they just reset the bit, so they'd have a clean test. All it required was some warm body to look at the test sequence as a whole, but no one had the time. Again, that single college kid might have saved the mission... but NASA was too cheap.

What concerns me is this: they're going to spend their time and money worrying over the hardware issues:

...Dr Pilcher said, that in the light of recent events, the timetable was wildly optimistic: "The jury is out on whether we have the technological capability."

rather than pay attention to managing what they've already developed. It's a bit like the aftermath of Challenger, where they went nuts on the hardware instead of looking at the fundamental problem, which was the prostitution of the program for political reasons. The outcome of that is that we now have a NASA which is completely paranoid about public opinion and afraid of its own shadow when it comes to safety, but which still won't look at the whole picture, and still twitches to the political beat.

It just really pisses me off! Pathfinder worked beautifully (despite a scary airbag system, which was what I figured would fail), and probably did so because of the long hours and very hard work everyone did; I know I did my share of 14-18 hour days on the little piece of it I had. It was so successful that NASA said, "Wow! That was really cheap! Let's see how much more we can cut out of the budget..."

So here we are: decent, low-cost hardware, and crappy, low-budget management. But guess which one is going to get the tarbrush?

---

Theres stuff we need to do before Mars...

[WolfWithoutAClause]

The problem with mars is that theres no obvious way to make money.

I think we should first mine Eros (that's a near earth asteroid.) Estimates indicate that it has 20 TRILLION dollars of ore on it- its 3% metal! It has everything, gold, plutonium, platinum...

There's nothing wrong with money. Money makes the satellites go around, and the sort of capabilities that you need to mine Eros will help get to mars- and probably pay for it.

And besides we need need to be able to stop the next dinosaur killer asteroid... living on Mars won't help much with that. Chucking around lumps of asteroid will.