"I have yet to meet my first independent safety researcher, who looks for flaws in life-critical equipment for fame or money or ego…"

You have now, this is what I do for a living.

"The article also completely ignores the fact that DO-178B certified code, as found embedded in civil aircrafts, has a perfect safety track record"

DO-178B is a process, and not a "certification". It's utilization has not produced a "perfect track record" in safety critical systems.

You should know better than to make such grossly incorrect statements. Experts in this industry know for a fact that testing only proves the lack of a "tested for" failure condition, and cannot be relied upon to assure safety.

DO-178B is but one of several criteria by which the Aircraft certification authorities judge an avionics system to be safe.

Software is too complex to be assured by testing alone.

Static analysis is merely a tool, and is not a solution to this problem. You should be well aware of the limitations of Static analysis.

Regulators no longer are provided with access to design and testing data sufficient to fully review a product for safety, and they routinely rely heavily on the manufacturers, thru delegated authority, to regulate themselves.

Design defects can only be eliminated in the application of a transparent review process. Complexity issues in "limited human cognition" and "human bias" are the key causes of life threatening defects in safety critical systems.

Without external review and oversight by "outside consultants" such as myself, life threatening safety defects routinely go undetected.

All safety critical systems will ultimately be "public domain" for this very reason. Just as proposed by the Software Freedom Law Center.

The FAA does not absolve aircraft manufacturers for liability, as the case with the FDA and Medtronic. (see Varig v. United States) 28 USC 2680(a) Discretionary function exception

DO-178B software does not "in fact" have a "perfect safety track record".

Turkish Airlines Flight 1951 was a passenger flight which crashed during landing to Amsterdam Schiphol Airport, Netherlands, on 25 February 2009, killing nine passengers and crew including all three pilots.

software defect utilized "invalid altitude data" as valid.

"Tests showed that the Rockwell Collins Enhanced Digital Flight Control System (EDFCS) uses radio altitude values that are characterised as ‘non computed’ (unusable,) whereas this characterisation should have prevented this."

http://en.wikipedia.org/wiki/Qantas_Flight_72

ADIRU software defect utilized "invalid data" as valid.

This device was certified to DO178-B Level A, the very highest level of certification, where failure can result in catastrophic loss of aircraft.

1 August 2005, Boeing 777-200 registered 9M-MRG, from Perth to Kuala Lumpur, Malaysia.

use of invalid data as valid

This device was certified to DO178-B Level A, the very highest level of certification, where failure can result in catastrophic loss of aircraft.

user:	aquila
created:	July 26, 2010
karma:	20