Steam And Software Flashcards
Computers and safety-critical systems
Scs- failure or malfunction could result in death, serve damage, environmental harm
More more computers are used to control dangerous devices
But need to ensure computers are introduced into SCS as soon as possible
“Anything that can be used to kill someone will be regulated”
James watt
Improved on newcomers designs
Improved efficiency patterns
Could be placed on ships and trains
Expired 1800
Higher pressure means more power but increased danger
Watt refused to increase pressure but when the patent expired other inventors brought new designs with frequent explosions
Problem?
Inferior materials used in production Built to low standards Mechanics lacked training Big problem with quality control Training was informal
Was a catch 22
Watt found that this technology he created was out of his control and increasingly dangerous, but if they refused to make it others would develop the technology many were untrained. Used influence to press for safer and better
Regulation
Not much was done because steam was so important
Many said the gains were worth the trade offs
Eventually numbers could not be ignored
Law in Us made it so had to have some training to use steam boat boilers
Uk in later 19th enforced high standards of design manufacture and operation
Parallel 1
Risks from steam engine came from boiler not steam engine
Boiler tech lagged behind improvements of engines
Computers now have potential for great economic gain
Computer hardware has advanced
But development of software engineering has been slower
Better to keep things simple
Dampen enthusiasm and confidence in computers
Use simple safe designs
Parallel 2
Little scientific understanding of boiler compared to engine
High pressure made design obsolete and exposed weakness in materials
Scientific foundations of software engineering are still being developed
Can use trial and error but very slow
There is both invention and science but has mostly stayed in invention
Need more in-depth understanding of whether inventions are effective
Use theoretical foundations for tests
Parallel 3
Safety features for boilers did not work (fuse plugs, safety release valves) because they were not based on scientific understanding
Software engineering deals with both abstract systems and concrete systems and should be evaluated as both
Start theoretical but eventually run on systems
Formal methods
Abstract ensuring safety
Concrete- need more validation of hypothesis supporting techniques
Parallel 4
Safety devices not work because of lack of scientific knowledge and narrow view to attempt to design solution without looking at other factors in environment
Social and organization
Airline expert system which caused quality to lower
Stopped taking responsibility and making own decisions
Removing faulty humans is not the solution
Must foresee all future scenarios and correctly determine response
Highly automated still need human supervision, maintenance, and observation
If not placing high priority on safety then best efforts from tech staff defeated
Parallel 5
Operators of steam engine received most blame but designers were without blame
Tendency to blame owners is still with us
Need to understand human psychology and behaviour
Take human factors seriously and involve human engineering experts
Parallel 6
Early steam had low standards of workmanship and lacked proper training
People lacking in-depth knowledge can be found building safety critical software systems
Standards and licensing must be carefully composed
Construct flexible standards