Talk:Safety-critical system

Compliance

Non of the Human made Systems are compliant to that error-rate not even Humans itself. Where Do You Got That From?

"A life-critical system is designed to lose less than one life per billion (10^9) hours of operation."

16:33, 16 February 2006 (UTC) Jan Girke

10⁹ hours of operation

Comments:

• 10⁹ hours of operation = 10^-9 failures per operational hour (according to failure rate, $MTBF={\frac {1}{\lambda }}\!$)
• The number can be found in e.g. hazard analysis and AC 25.1309-1A.
• Software can be designed for this probability area with DO-178B, level A (according to AC 20-115B).
• Hardware will typically need redundancy for this low failure rate.
• The number is also probably linked to humans in aviation with something like Pilot certification in the United States#Medical Certification and Requirements and requirement of co-pilots in commercial/large airplanes (software is my "field", not humans -but this should be obvious).
• This number in requirements (from FAA with e.g. FAR 25 §1309) for software and hardware (referred to as "system") with "catastrophic" failure conditions (ref e.g. hazard analysis) is the main reason for the high cost of (at least some) avionics.
• I have not contributed to 10⁹ products myself, but I have made software for 10⁷ products which when integrated as part of a larger system (with redundancy) becomes better than 10⁹.
• Did you know the rubber on the windscreen wiper has been a major pain in the ... for a large airplane manufacturer? Physics sucks big time when working with failure rate requirements in this area.

With all these comments, the number may be misplaced. The article could focus on "higher level stuff" (but then again, the number has major impact on the design of life critical systems). Anybody up for a vote?

Nordby73 22:48, 16 February 2006 (UTC)

Request for expansion

I'm interested in reliable system design - software design in particular - and I thought that article could be expanded. It explains well what a life-critical system is, but says fairly little about how such systems are designed and implemented. There was in fact no direct mention of Reliable system design and Redundancy (engineering), so I added these under the "See also" section. Also: reliability regimes are only one aspect of reliability; it would still be possible to create a reliable system that is not fail-operational, fail-safe nor fault-tolerant. (Think about cable cars: what happens when a cable fails?)

This is not to be construed as a criticism, because the article is already good in what it covers, and things I find missing are related to reliability in general, and are not particular to life-support systems. So, this is all just a suggestion... GregorB 20:58, 17 February 2006 (UTC)

Proposed merge with Safety-involved systems

Minimal article that discusses a strongly related subject. Could be treated as a near-synonym in the lead paragraph, but I'm hesitant to do the merge myself as I don't know much about safety-critical systems. QVVERTYVS (hm?) 18:21, 23 December 2013 (UTC) Done

Name?

If this is "more commonly a safety-critical system", why not just rename it to safety-critical system? Disposable Redshirt (talk) 02:06, 30 March 2017 (UTC)

@Disposable Redshirt: That's exactly why I placed the speedy deletion tag that you removed - so the redirect at Safety-critical system could be deleted to allow this page to be moved there. – Train2104 (t • c) 02:09, 30 March 2017 (UTC)