Behavior Pattern? F-35 Software Lawsuit
Stephen Trimble has reported that there is a lawsuit filed against Lockheed Martin claiming that the F-35 software is dangerous.
A newly unsealed lawsuit accuses Lockheed Martin of developing corrupt and possibly dangerous software for the F-35 Joint Strike Fighter flight control system and then lying about it to the government.
Sylvester Davis, former software lead and software product manager for the F-35 flight control application at Lockheed Martin, has filed the False Claims Act suit in US District Court for the Virgin Islands.
Davis’ lawsuit recommends to the court that Lockheed should “immediately” stop developing software for the F-35 to “avoid further waste” of resources and the “serious risks” to F-35 pilots.
“The software contains substantial corruption,” says the lawsuit, “which has multiplied significantly the risks that the software will not operate as intended.”
This is especially important because F-35 program officials have stated that software is on track. The F-35 has more than three times the amount of software than an F-22. The F-22 was plagued with software problems in its development. The F-35 program has stated more than once that they have significant lessons-learned from the F-22 program in the area of computers and software.
This is not the first lawsuit against Lockheed Martin on fighter aircraft issues. Recently a lawsuit appeared stating that the F-22 program ignored engineers on low observable quality concerns.
Are these real concerns? Are they true? Is this a behavior pattern by a company? Should we worry?
Since its a false claims case Davis may be awarded 15-30% of whatever Lockheed Martin coughs up if they lose.
http://en.wikipedia.org/wiki/False_Claims_Act
Software “was” a problem with the F-22? Software is a problem with the F-22. Lockmart is still chasing bugs.
Yeah, the F-35 is right on track, sure.
A couple things to consider:
1. All systems are run by software, either code that runs in a CPU, or hard coded ICs.
2. The F-22 had such a problem in the software area because all the systems ran in a common environment, think single OS. So, when a problem occurs, it can freeze the whole system, think BSOD.
3. The F-35 runs each system in it’s own environment so that a problem cannot bring the whole system down.
4. The F-35’s software is in C+ which makes programming and debugging vastly superior to the F-22’s system.
5. The lawsuit does not say the software IS defective, but that the safety protocols were not followed that COULD result in problems down the road.
6. The lawsuit claims that each entry, in each time card, is a separate violation to the “False Claims Act”. This is a clear attempt to get as much money as possible. If calculated this way, the damage award would be dozens of times higher than the entire software development cost.
7. The lawsuit claims “First delivery of operational aircraft was anticipated for fiscal 2008. That delivery did not occur as promised, due, in large part, to the issues raised in this complaint.”
It is common knowledge that the SWAT program is responsible for the delay.
Actually if you re-read the lawsuit claim, the language is more serious than just the wave away by an F-35 enthusiast.
I did read it and I know the allegations of falsified docs stating SEAL1 compliance, etc.
I was not doing a “wave away”.
My points still stand that there is no claim that the software IS DEFECTIVE, just that safety protocols were not followed in it’s development.
btw, if the lawyer is in Texas, why file the claim in the US Virgin Islands? Judge Shopping?
Interesting.
So claims of not following safety standards for software and claims of software “corruption” (in the software sense)… don’t define claims of defect.
As to your later part, I don’t know. I am sure that you can’t believe otherwise.
There was no claim that the software was corrupt in any specific sense, just that by it not following the safety protocols, it MUST be corrupted.
If it is found out they did fraudulently deceive the Gov, then they should be seriously fined.
I just find it very dubious that they went with the USVI as their court of choice. The pessimist in me has a few ideas:
1. The farther away it is, the higher the costs for LM and the likelier they will just settle.
2. The USVI has a lower technology base thereby making it easier to confuse a jury into believing guilt.
In the legal world, Federal Courts are considered the top of the ladder, wither in the Virgin Islands, or anywhere else. One Attorney friend liked to tell me that Federal Courts are considered “REAL COURT”.
Attorneys often use a court that they are more familiar with, and that is more likely the motivation for the case being located in the Virgin Islands vs. some other location.
Sorry Spudman, the whole story of this court not understanding technology vs. any other Federal Court is just pure spin. The same national caliber expert witnesses will end up testifying in this district, just as they would anywhere else in the US.
Since the law firm is based in Texas, I doubt he is “familiar” with the USVI court to the extent you suggest.
I did not say it was the Court’s inability to understand tech, but the jury’s.
Since this is a jury trial, it’s the competence of the jury that is the issue. There is just no reason, outside some perceived advantage it gives the plaintiff, to use the USVI court.
The first thing LM should do is ask for a change of venue.
Large national legal firms have certain districts that they prefer, this is a well known fact. What state a firm comes from has no relevance in this. As someone who has testified in Federal Court many times in different places in the US, I can assure you that both the judge and the jury will be more then capable of dealing with the facts.
We should not deviate from the original topic here. This is one more example of the continuing saga of controversy surrounding F-35 fighter program. I certainly hope that the end product is a safe and competent fighter. With the political witch hunt that is burning the F-22 at the stake, LM is not going to continue to get unlimited chances to make the JSF successful.
Mr. Palmer please contact me.
This is a complaint, which is just someone making a claim that he hopes to prove at trial. As with the F-22 complaint, it’s worthless to speculate about whether or not the claims made are true before there’s any evidence presented.
Royce: You are right, this is just an unproven complaint at this point.
However, if you look at this as a piece of the overall picture:
1.The large amount of code being written for the JSF (far more then any other fighter in history).
2.The lack of flight testing that has occurred up to this point.
3.And now a legal complaint from a software engineer from the program who had has indicated that the code is dangerous.
It certainly looks suspicious and only time will tell. Flight testing results will be interesting to watch and should give us a better idea if there is truly something wrong.
#1 There is also has been the largest number of lab hours to compensate. Throw in things like DAS that have not existed, the amount of code climbs.
#2 If we were in a MYB (Multi-Year Buy) this would be a problem, but we are still in SDD and testing continues. 2009 will see a large jump in flight tests and 2010 will be a phenominal jump in flight tests.
#3 Complaint states that code COULD be dangerous, not IS.
The overriding thing that LOOKS suspicious is the fact they filed in USVI considering all the parties are in Texas.
To SpudmanWP:
Do you have any links or sources regarding the F35 that specifies each system runs in its own environment?
Combine these two thoughts together and see where they lead:
1. Security protocols around details concerning the F-35 and its development, supplier chain have already been breached and that LockMart has already admitted this to be the case (while simultaneously claiming that no “secure” information was stolen.
Point: the F-35 program has already been hacked into
2. A complaint alleging that LockMart failed to rigorously supervise the quality controls of its code base surfaces, which if true may mean that portions of the code base for the F-35 can/have been rendered corrupt (in the coding sense) and thus endangering the lives and safety of the pilots who will fly the plane.
Point: the F-35 program may be nowhere near as secure/stable/safe as the powerpoint warriors would love for anyone to believe
Combine points 1 and 2 … and what do you get?
For a technical slant:
http://www.embeddedstar.com/press/content/2004/1/embedded12564.html
Being very familiar with this stuff, although not with the JSF program, I don’t see any merit to these claims. (For the record I am not affiliated with the Mathworks or any aerospace company at the moment.)
Probably the wrong audience for this, but whatever.
The lawsuit is quite straightforward in describing what happened. This was an engineering catfight over whether to use the Matlab Real-Time Workshop from the article above (conspicuously absent in the deposition) or to code things by hand. Davis thinks coding things by hand is the right way because this is how you implement whatever SEI standards the PMs sold the customer on. The truth is that the RTW is adding a layer of complexity to the compiler level, not to the software development process itself, which is all SEI addresses.
The fundamental problem with applying SEI standards to flight controls development is that you’re trying to certify a very complex mathematical algorithm by checking boxes on some standard that doesn’t have a straightforward interpretation in this scenario. The best you can do is to certify that something behaves as expected at various points within a feasible operating window. You might not like it, but this is how in-flight validation is done as well. Every single plane you’ve ever flown on has had to deal with similar issues. LM’s problem was that they signed up for a standard that didn’t explicitly recognize this.
Addressing some of the technical comments here:
- Spudman is right about the claims saying absolutely nothing about the software not functioning correctly. It goes out of its way to make sure this is never claimed.
- Spudman is wrong in arguing that [C++] makes the software development process somehow more safe. It’s a much more complicated language, and programming and debugging is a lot more difficult. But it’s a lot more powerful and necessary for more complex programs.
- Insinuating that the software could have somehow been “hacked” from outside LM is ludicrous.
I agree with your assessment regarding C++ especially with regards to memory addressing that can lead to issues such as buffer overruns.
One complaint that I found interesting in the lawsuit is the alleged mixing of SEAL 1 and non SEAL 1 code in the same address space. Modern operating systems typically protect the memory spaces of different programs to prevent potential corruption so one program crashing should not affect another concurrently running program. Additional measures include not running application code within the OS kernel, I.e. ring 0, so that bad code will not crash the OS.
Without knowing what levels of quality control and rigor that is implied by SEAL 1, the alleged corruption of code is difficult to quantify and not necessarily an indicator of potential software defects.
punctilious post. upright one decimal where I quarrel with it. I am emailing you in detail.
“I agree with your assessment regarding C++ especially with regards to memory addressing that can lead to issues such as buffer overruns.”
ONLY if you write a C program in C++. C++ is a safe language if you use it properly, you have the standard library which eliminates 95% of your need for pointer arithmetic (the reason for issues such buffer overruns etc.) of course only if you use it. Another fact is that, one of the consultants on the JSF (F-35) project on coding standards (one of the important fundamental ways to achieve design and code quality) was actually Bjarne Stroustrup (the inventor of C++) known for extremely high standards of quality and very little tolerance for bullshit. The probability that the F-35 software development is the best ever in DoD projects are very high, they use a real world industry strength robust technology proven in many different areas including space explorations and all of the most complex systems ever built.