Errors, Failures, and Risks
Nishat Tasnim Shishir
Lecturer,
Department of Computer Science and Engineering,
Daffodil International University.
When Computers Fail: Errors That Impact Lives
Computer errors can cause wrong bills, false arrests, and even
dangerous accidents. Many mistakes come from poor design,
carelessness, or outdated information.
To prevent major failures, businesses and users must double-check
data, update systems, and use computers responsibly. While no system
is perfect, careful management can reduce risks.
Common Billing and Database Errors and Their Impact on Individuals
Incorrect Billing: Mistakes such as a woman receiving a $6.3
million electricity bill instead of $63, or the IRS sending
outrageous tax bills due to system errors.
Database Errors: Issues like credit bureaus mistakenly listing
people as tax delinquents or individuals being wrongfully labeled
as criminals due to incorrect data in law enforcement databases.
Consequences of Mistakes: These errors have led to denied loans,
wrongful arrests, and even harm to innocent people due to
misinterpreted information or system failures.
8.1.3 System Failures
Importance of Computer Systems in Modern Society
o Dependence on computers for communication, power,
medical, financial, retail, and transportation systems.
Examples of Failures and Causes
o BlackBerry Failure: Nine-hour email outage due to faulty
software update.
o AT&T Telephone Service Failure: Software error in a 4-
million-line program causing disruption.
o Telecommunications Switching Program: A typo in a 2-
million-line program leading to network failures in major
cities.
o American Express System Failure: During the holiday
season, verification system failure led to overwhelmed call
centers.
o Skype Login Overload: Surge of log-ins after Windows
update caused a two-day outage.
o Galaxy IV Satellite Failure: Caused wide-ranging disruptions
like pager service and flight delays.
Stock Market and Financial System Failures
o Tokyo Stock Exchange Crash: Error in software upgrade led
to shutdown.
o Charles Schwab System Crash: Two-hour outage due to
software glitch.
o London Stock Exchange Crash: A malfunction that lasted
almost 8 hours on the last tax day.
Transportation and Reservation Failures
o Amtrak Reservation System Failure: Caused delays during
Thanksgiving weekend.
o Virgin America Reservation System Failure: Problems with
website and check-in kiosks post system switch.
Mars Climate Orbiter Loss
o Miscommunication between teams using different
measurement systems (English vs. Metric) led to the failure
of the $125 million project.
Inventory Management System Failure
o A modified system caused significant errors in inventory,
pricing, and order processing, leading to financial losses and
poor service.
Voting System Failures
Problems with Electronic Voting Systems
o Some systems experienced crashes, failure to count votes,
and programming errors that resulted in incorrect vote
tallies.
Security and Integrity Issues
o Security flaws in electronic voting systems allowed potential
manipulation of results.
o Insufficient training for programmers and election officials
led to vulnerabilities.
The 2000 U.S. Presidential Election and the Help America Vote
Act
o Issues with paper and punch-card ballots led to reforms.
o The introduction of electronic voting systems introduced
new problems, including uncounted votes and incorrect
votes.
Lessons Learned from Failures
Lack of Adequate Planning and Testing
o Many failures stemmed from insufficient testing, security
concerns, and a lack of thorough planning.
Importance of Backup Systems
o Systems must have fail-safes, especially for critical
infrastructure like telecommunications and voting systems.
Honesty and Transparency in Error Management
o Addressing errors quickly and honestly is vital to avoid
significant consequences, as seen in the CTB/McGraw-Hill
standardized test error.
Abandoned Systems
Some computer systems fail so badly that they are abandoned,
wasting millions or even billions of dollars.
Examples include:
o A British food retailer spent $500 million on an automated
supply system that failed.
o Ford abandoned a $400 million purchasing system.
o California and Washington state motor vehicle departments
spent millions on systems that were eventually abandoned.
Common causes of failure:
o Lack of clear goals and specifications
o Poor management and communication
o Unrealistic estimates for time, budget, and resources
o Use of untested or unfamiliar technology
o Refusal to admit problems
Legacy Systems
Legacy systems are outdated but still in use, often due to high
costs or disruptions caused by a complete overhaul.
Issues with legacy systems:
o Hard-to-find replacement parts for old hardware
o Outdated software running on modern hardware
o Programmers no longer familiar with old programming
languages
o Lack of proper documentation and design resources
Examples:
o US Airways' merging of old reservation systems, leading to
service issues.
o Old airline systems designed for mainframe computers of
the 1960s-1970s.
The need for good documentation, design flexibility, and system
upgrades.
What Goes Wrong?
Computer systems fail for two main reasons:
1. The job is inherently difficult (complex systems with real-
world interactions).
2. The job is poorly done (inadequate design, development,
and management).
Common failure factors:
o Design and development:
Lack of safety risk attention
Incompatibility between software and hardware
Poor testing and insufficient planning
o Management and use:
Data-entry errors
Inadequate training and failure to update databases
Overconfidence in software
Overconfidence in Software
Overconfidence about system reliability can lead to problems.
Unrealistic estimates and lack of understanding of risks can cause
failures in critical systems.
Examples:
o Failure to account for real-world issues (e.g., physical
interference like spilled coffee in an airplane cockpit).
o Pressure to hide flaws or misrepresent safety and reliability
for business or political reasons.
Reuse of Software: The Ariane 5 Rocket and "No Fly" Lists
Ariane 5 Rocket Failure: A software error during launch caused by
using software designed for the older Ariane 4 rocket. The
software could not handle higher velocity calculations, leading to
a failure.
"No Fly" List Mistakes: Outdated software caused false matches
on terrorist watchlists, mistakenly flagging innocent people as
security threats.
Key Lesson: Reusing software can be beneficial but requires
careful reevaluation, testing, and adjustments for new
environments or purposes.
