1.1 Reliability and Integrity

Exercise 3-2

The following materials provide more details on the Therac-25 case.

• ComputingCases: Therac-25
• An Investigation of the Therac-25 Accidents
• Case Analysis: Therac-25

Computer Reliability Exercise 1

The articles below cover different types of failure in computer controlled systems. A useful classroom task is to analyse these articles and consider how significant hardware failure, software failure, user failure, and corrupt data are in causing the problem.

• Radiation Offers New Cures, and Ways to Do Harm (NYTimes)
• NYC Fire Department tech in the hot seat as firefighters complain about being sent to wrong addresses (ScientificAmerican)

RISKS and Risk Factor are two excellent resources frequently updated with reports of reliability and other ITGS related computing issues.

Computer Reliability Exercise 2

Research the following incidents. In each case, to what extent can the system software be blamed for causing the problem? What do these accidents tell us about the issue of software reliability?

• Accidents involving the Airbus A320 aircraft, including Air Inter Flight ITF148 to Strasbourg in 1992 and Lufthansa Flight 2904 to Warsaw in 1993.
• The accidental shootdown of Iran Air Flight 655 by the USS Vincennes in 1988.
• Boeing Promised Pilots a 737 Software Fix Last Year, but They�re Still Waiting

Cars and software reliability

Modern cars contain numerous microprocessors and thousands of lines of computer code. Software reliability problems have their own section here because of their ubiquity and their potential to be extremely dangerous (although it must be said that most issues are not life-threatening).

• In January 2014 GM Recalls 370 000 Pickup Trucks for Software Update to Reduce Fire Risk
• In November 2013 Honda recalled 344,000 vans with a software glitch in the braking system
• In June 2013 Chrysler recalled 600,000 vehicles because of software problems that could prevent seatbelts and airbags from deploying correctly, and again for another fault in March 2014.
• Chrysler recalled 14,800 vehicles to recalibrate transmission software after complaints of stalling.
• Toyota recalled 436,000 vehicles in 2010 after discovering a bug in the cars' braking software.
• In December 2012 Ford recalled 90,000 vehicles are a software fault in the cooling system increased the risk of an engine fire.
• In October 2011 Jaguar had to recall almost 20,000 cars after a bug meant disabling cruise control 'in the normal manner' was difficult or impossible.

Computer Reliability: Examples for Further Reading

The ten biggest computer failures of 2012 (Telegraph) is a slideshow of the worst computer failures during 2012. Apple Maps glitch could be deadly reports on some of the (many) errors found in Apple Maps when it was released in 2012.

Post Office wants to get to bottom of IT system allegations A quite frightening article about how Post Office workers have been disciplined and even sacked for fraud, despite allegations that accounting errors are caused by faulty software. In July 2013, an investigation found hardware and software problems in the Post Office's systems which may have been responsible for the errors ( ComputerWeekly article, BBC article).

Google Spreadsheet, Excel Web App 'cannot be trusted' examines the potential impacts of unreliabile free software - in this case, Google Doc's Spreadsheet application and Microsoft's Excel Web App, which failed a variety of mathematical tests involving common spreadsheet functionality.

A software bug in the US military's Patriot Missile system led to an inaccuracy that increased over time that ultimately led to the death of 28 US servicemen. The linked article gives a very clear explanation of how such a simple error can cause significant problems (the missile's timer was out by less than half a second - but when a missile is travelling at 1.7 km/s, that equates to over half a kilometre off-target).

User and Computer-related Errors Involved in Two Australian Aircraft Incidents (IEEE) describes two 'computer' errors airboard passenger aircraft. The first caused an Airbus A330 to unexpectedly maneuver without the pilot's command; the second involved a simple data-entry error that could have resulted in loss of the aircraft.

Using a Software Bug to Win Video Poker (Wired) describes how a man exploited a software bug in a gambling machine to win thousands of dollars.

Chrysler to recall 840,000 vehicles because a software error causes the wrong airbags to be deployed in a crash.

US State Governments Can't Shake IT Woes gives a good overview of a multiple of IT systems failures in the US government.

20,000 traffic tickets wrongly issued is a good example of a failure in automated systems.

RISKS digest is an excellent and very active mailing list of "risks", many (but not all) of which are related to failures in computer systems.

Dove Evolution and the Digital Manipulation of Models

Digital manipulation of models' photographs is a commonly discussed example. In 2006 Unilever launched its Dove Campaign for Real Beauty, designed to highlight natural beauty. The advertisement (video) shows the transformation of a model using both make-up and digital computer manipulation. It makes an interesting lesson starter when teaching digital manipulation of models. Can you believe your eyes in the digital world? is a BBC article and video that deals specifically with the digital manipulation of models' images and the possible impacts. A mascara advert featuring Natalie Portman was banned after it was realised images had been manipulated in Photoshop.

On the same topic, in 2009 French MPs proposed a law that would require 'health warnings' on any advertising images that had been digitally manipulated. This proposal became law in late 2017, requiring images that had been digitally manipulated to carry a label to that effect.

Finally, this fun link shows celebrities 'Photoshopped' to look like ordinary people.

Digital image manipulation relates closely to the IB Theory of Knowledge (TOK) course.

Digital Manipulation Ethics - Unhate Campaign

In 2011 fashion label The United Colors of Bennetton launched its 'Unhate' campaign and gathered immediate controversy. The campaign - aimed at reducing the 'culture of hatred' around the world - featured manipulated photos of world leaders kissing. Examples include a photo (which was later pulled) showing the Pope kissing Egypt's Sheikh al-Azhar Ahmed al-Tayeb. Another featured Barack Obama and Venezuelan President Hugo Chavez.

The photos can be viewed here (the image of the Pope was pulled but can easily be found elsewhere). They are excellent examples of how far digital manipulation can be taken, and with a mature class can be used as a good basis for a discussion of digital editing ethics.

Digital image manipulation relates closely to the IB Theory of Knowledge (TOK) course.

Famous Image Manipulations

Myriad websites contain examples of "the best" (or worst) digital manipulations: some of the more notable examples include The Guardian's When Photoshop can get you into Trouble, Iran 'faked missile test image',

 When Photoshop can get you in trouble (Guardian) On The Economist's Cover, Only a Part of the Picture (NYTimes) French MPs want health warnings on airbrushed photographs (Telegraph)

Digital manipulation can also have significant legal impacts: in a very worrying case, the Guardian reports that the Metropolitan Police deliberately manipulated a photograph submitted as evidence in court, in order to exonerate themselves in the accidental killing of civilian Jean Charles de Menezes.

Digital image manipulation relates closely to the IB Theory of Knowledge (TOK) course.

Spotting digital manipulations

The Fake or Real? gallery from PBS and Photo Tampering throughout History are good starter resources for this topic. The article Digital Forensics - 5 ways to spot a fake (Scientific American) is another essential read.

This PBS Nova video describes new software which uses mathematical techniques to look for tell-tale signs of image manipulation (such as cloned regions). Four and Six is a company that produces software to authenticate digital images, and identify techniques commonly used by digital forgers. Their website contains a lot of detailed information about the processes they use. Similar techniques are also used in art authentication (PBS).

Database integrity issues

The lack of data integrity can cause serious social impacts. While a 'computer glitch' blamed for a 12 quadrillion euro telephone bill may be easy to spot and even quite humourous, other impacts are more serious.

The Transport Security Administration (TSA) is notorious for errors on its terrorist watch list, which has led to many passengers being inconvenienced, including an  Eight-Year-Old boy (Wired) called Mikey (NYTimes).

Elsewhere, Dead girl given truancy warning (BBC) and Outrage at 500,000 DNA database mistakes (Telegraph) highlight the risks of incorrect or out of date data. In the infamous 2000 US election battle between Al Gore and George W Bush, the integrity of a database of convicted felons (who should be unable to vote) was called into question, potentially disenfranchising many voters.

Electronic Medical Records (EMR) & Electronic Health Records (EHR)

Electronic medical records resources:

• Two resources from the New York Times, Digital Health Records: The Hard Road Ahead and The Ups and Downs of Electronic Medical Records, provide an excellent and very detailed insight into EMR, their benefits, and their potential problems.
• Hospital puts medical records snapshot on smart cards (Network World)
• Learn about e-prescriptions
• Maui Family Doctor Achieves High Efficiencies With EMR Good example of EMR in a small clinic (video)
• Microsoft's HealthVault Example online EMR.
• Hunt wants paperless NHS by 2018 given the NHS' infamous history with IT projects, perhaps this article could also go in the IT Systems and Organisations page under 'project failures'.
• It is patently obvious the NHS must have online patient records (Computer Weekly)

Electronic Voting software and lesson plan

Electronic voting is a controversial topic. This E-Voting lesson plan uses a simple Java application I wrote to simulate the e-voting process. Students get to vote and then are presented with three sets of results - two of which are falsified. This is a useful practise exercise to stimulate discussions about e-voting and the potential problems that may arise.

The New York Times article Voting Test Falls Victim to Hackers and the video Why Internet-Based Voting Is a Bad Idea are also useful for this task, as are the articles below.

Electronic Voting articles

E-Voting impacts and issues

Report: Voting Machine Errors Highlight Urgent Need for U.S. Database (Wired) describes many, many problems that have occurred with e-voting machines in recent years. Some of them are quite unusual.  E-voting system awards election to wrong candidates in Florida (ComputerWorld) and Voting Out E-Voting Machines (TIME) both detail further problems.

Oscar's E-Voting Problems Worse Than Feared analyses the problems that faced e-voting systems designed to vote for Oscar nominees, while 'Fake votes' cast in France's first digital election (BBC) explores France's June 2013 open primary mayoral election - both articles are a stark reminder of the myriad problems facing such systems.

Finally,  this is a letter to President Obama about e-voting, written by elections officers and computer security experts - and urging him to resist calls for Internet voting. Online voting is impossible to secure examines the various security issues related to online voting, and discusses why voting is much easier to attack than other secure applications such as online banking.

Solutions

Science Daily's 'Voter-Verifiable' Voting System Ensures Accuracy And Privacy explains how paper-trails are needed on voting machines, while Aussies Do It Right: E-Voting (Wired) discusses another possible solution - open source voting software (this is a good article for students who believe open source software is "less secure".

Police use of IT: Offender databases

Online offenders databases remain a controversial topic, with security, privacy, and integrity being key issues. Nevertheless, many such databases exists, especially in the US: Family Watchdog lists details of sex offenders living in the community (US), while the Michigan Public Sex Offender Registry (PSOR) contains records of sex offenders in the state of Michigan. Florida Department of Corrections Offender Database has online records of prison inmates, released inmates, and fugitives. The Sensible Sentencing Trust is a similar database of offenders in New Zealand - interestingly this is not operated by the government, which could raise further issues related to privacy and integrity.

Mugged by a Mug Shot Online (NY Times) discusses some of the potential long term ramifications of exposing such data.

Police use of IT: GPS offender monitoring and tracking

Prison without Walls is an excellent article about the use of GPS trackers to monitor paroled offenders. It covers the technical aspects of the system (Strand 3: IT Systems) plus the plethora of impacts and issues (Strand 1), including economic impacts and psychological impacts. However, according to the LA Times and The Atlantic, some of these ankle GPS trackers have serious reliability and security issues which mean prisoners have been able to disable them.

SecureAlert, GPS Monitoring Solutions, and G4S are all companies that develop and supply offender tracking technology - their home pages include a lot of detail and examples about how the systems work and their impacts.

This topic also makes a great practice paper 2 news article.

The Net

Amazon.com | Amazon.co.uk | Instant Video

An old film (it features floppy disks!), but a good one for ITGS discussions. Sandra Bullock finds herself "wiped from existence" and given a false criminal record after stumbling into a criminal gang. This is a good film for discussing databases, ITGS issues such as data integrity, privacy, and security, and the increased trust that many in society place in computer systems.

Driverless cars and ethics

Driverless or self-driving vehicles are often promoted as being safer than human drivers. However, there may be situations in which an accident is unavoidable. In these situations, how should a driverless vehicle be programmed to behave? Which course of action should it take if all have negative outcomes? And, of course, who takes responsibility for any damage that is caused?

This is a topic which links to ITGS and TOK. The ethical dilemma of self-driving cars (video) is a good introduction. Why Self-Driving Cars Must Be Programmed to Kill and Ethics of Self-Driving Cars are great articles that examine the topic in more detail.

Uber driverless car accident

In March 2018 an accident occured which was reportedly the first death caused by a driverless vehicle. The Uber self-driving car hit and killed Elaine Herzberg, 49, in Arizona. The human monitor in the car also failed to spot the pedestrian until seconds before the collision. Uber stopped all self driving experiments in the aftermath of the crash.

Velodyne, the company that produces the sensors for the cars, reported that the sensors were working correctly - suggesting a software issue may have been the cause. It was later reported that the car's sensors detected Herzberg, but chose not to swerve as it was uncertain about the nature of the obstacle.