Todos
← Back to Squawk list
Boeing discovers more software issues in the Max
The United States planemaker confirmed to Reuters news agency on Tuesday that one issue involves hypothetical faults in the flight control computer microprocessor, which could potentially lead to a loss of control known as a runaway stabiliser, while the other issue could potentially lead to disengagement of the autopilot feature during final approach. Boeing said the software updates will address both issues. (www.aljazeera.com) Más...Sort type: [Top] [Newest]
As a person who has worked in the information technology industry for 40 years, I can say with certainty that no software will ever be absolutely correct. Any company that’s worth their money will be constantly testing their software and they WILL find bugs. It will never end. It’s the reason why one gets patches for any software on a regular basis. This is not a lack of quality control. This is the abundance of quality control. EVERY company that writes software does this. If a software company does not do this they will go out of business.
As a software architect, designer, developer, tester and QA adviser, I see W2bsa's comment and raise you. There are no 'fail-safe' hardware or software solutions, only ones less likely to fail than others. The closest we can reasonably get in 2020 is highly reliable designs (heaviest, most cost, least performance) in fully redundant configurations. Never gonna happen, not in flight and not in business. Balancing acceptable risk (losses) is the key; public tolerance for how often and how deadly is ever changing. Flying is dangerous, but fly we must.
Steven Austin, agreed. Aircraft design has always been a compromise with cost, acceptable risk, and safety. This thought process is nothing new. Humans cannot make a perfect machine. Whether traveling by trains, planes, or automobiles, their will always be risk.
w2bsa, Stephen, and Thomas: I call. I have a similar 40 year background in software engineering, with the last 25 years spent in safety-significant software systems for medical devices. All of your general comments on software design are correct -- software of any complexity will have faults and defects. What alarms me (pun intended) is that, in the FINAL verification and validation phase of this software development, Boeing is still finding defects in the avionics software that have the HIGHEST risk severity possible: loss of the aircraft. In my experience, finding such new show-stopping high-risk defects at this late stage of development is a huge red flag that indicates 1) the software architecture is flawed in its ability to detect, contain, and mitigate high-risk failures ("exception handling"), and 2) the software development PROCESS employed was itself flawed in its ability to find and mitigate to safe levels these high-risk defects("defect containment") during review, inspection, and testing activities in the earlier requirements, design input, and design output development phases of this software. Finding high-risk failures in the avionics software at this late stage of verification and validation is truly alarming.
I have a family member who is a captain for Southwest. He loves flying the Max, touts the numerous safety feature improvements over previous 737 models, and "would take his family up in a Max tomorrow", if it were flying. Interestingly, that endorsement is exactly one of the litmus tests that we also use in medical device safety risk assessment: "would you connect your kids to this medical device?"
Sadly, I won't be joining him on a Max anytime soon. I've read everything I can get my hands about the Max software design and the development process employed in its creation. The deficiencies in both are truly scary to me. When and if it returns to service, I hope the Max has a long, accident-free service life, but I'll be sitting on the sidelines for a while yet.
I have a family member who is a captain for Southwest. He loves flying the Max, touts the numerous safety feature improvements over previous 737 models, and "would take his family up in a Max tomorrow", if it were flying. Interestingly, that endorsement is exactly one of the litmus tests that we also use in medical device safety risk assessment: "would you connect your kids to this medical device?"
Sadly, I won't be joining him on a Max anytime soon. I've read everything I can get my hands about the Max software design and the development process employed in its creation. The deficiencies in both are truly scary to me. When and if it returns to service, I hope the Max has a long, accident-free service life, but I'll be sitting on the sidelines for a while yet.
So by reading what you find on cnn or reuters or the seattle times, you feel that you have a better grasp of the situation than he does?
I never read any of your listed biased, trash rags. I just responded to Warren Craycroft's opinion. Don't put words in my mouth.
He was responding to Warren, not you.
I second Mr. Craycroft's input... I have some of the same software development background, and V&V. In my opinion, if Boeing was still run by engineers instead of MBAs, this high number of aircraft-loss-level software errors would never have made it into the air. If the "new Boinnggg" spent any money on software DERs, I hope those folks enjoy their new careers, whatever they will be....