Software Correctness

Question 1

What is the main structure of the essay? (10)

Answer 1

• Introduction - what is buggy software
• Why it’s important - implications
○ Medical/nuclear - fatal (Therac-25 1986-87)
○ Embedded systems - finance (Intel Pentium 1994)
○ High risk/complex science - robust (Mars Orbiter 1998)
• How to combat buggy software - preventions
○ Good software practice
○ Fault tolerance systems
○ Formal verification
○ Programming language innovations

Question 2

What are implications of buggy software? (3)

Answer 2

• In safety-critical systems, such as nuclear reactor controllers or medical software, errors could be fatal
• In embedded and commerical systems, such as processors, errors could have financial impact due to being recalled
• In high risk/complex systems, such as scientific experiments, non robust software could be

Question 3

Expand on the first implication

Answer 3

If software is incorrect or ‘buggy’ in a safety-critical system, such as a nuclear reactor controller or medicals surgery system, then a software error could cause loss of human life

Question 4

Give an example of the first implication (2)

Answer 4

• Therac-25, a radiation therapy machine that killed 4 and injured 2 people in 1986-1987
• Exposed patients to massive overdoses of radiation; due to the controlling software that had bugs in it, and no hardware interlocks or fault tolerance to deal with the software bugs.

Question 5

Expand on the second implication

Answer 5

Software errors could also have financial consequences, such as embedded systems being recalled or compensation for commercial sold systems injuring and failing.

Question 6

Give an example of the second implication (2)

Answer 6

• Intel Pentium processor bug 1994, a bug affected the FPU and cost Intel around $400 million to recall and fix.
• The bug caused the processor to return incorrect decimal results when dividing a number.

Question 7

Expand on the third implication

Answer 7

High risk and cost complex systems require very robust software, such as expensive scientific experiments.

Question 8

Give an example of the third implication (2)

Answer 8

• Mars Climate Orbiter 1999, communication with the probe was lost when a software bug caused an error in a momentum adjustment.
• The software outputted the calculated the momentum adjustment in the wrong units, causing the probe to fly too close to the planet and disintegrated.

Question 9

What are the ways to combat buggy software? (4)

Answer 9

• a. Good software practice
• b. Fault tolerance systems
• c. Formal verification
• d. Programming language innovations

Question 10

a. What does good software practice mean?

Answer 10

Conforming to common rules of software development to improve the quality of software code

Question 11

a. What is an advantage & disadvantage of good software practice? (2)

Answer 11

• Used to build up the integrity of software and ensure robustness and reliability
• Takes more time which could otherwise be spent implementing new features

Question 12

a. What are two approaches to good software practice? (2)

Answer 12

• Test Driven Development (TDD)

* Code Review

Question 13

a. What is Test Driven Development?

Answer 13

Writing tests before you write just enough production code to fulfill that test and refactoring

Question 14

a. What is Code Review?

Answer 14

A systematic examination of source code, where a developer walks through the code

Question 15

b. What are Fault Tolerance Systems?

Answer 15

Systems that can continue to function in presence of failures or faults in its hardware or software.

Question 16

b. What is an advantage & disadvantage of Fault Tolerance Systems? (2)

Answer 16

• Can ensure the running of a safety critical or expensive system by implementing fail safes.
• Economic cost as well as additional space and weight required, can be impractical for systems such as space ships

Question 17

b. What are two approaches to Fault Tolerance Systems? (2)

Answer 17

• Erlang

* Chaos Monkey

Question 18

b. What is Erlang? (2)

Answer 18

• A programming language designed for developing robust and reliable programs for servers and embedded systems.
• The error handling mechanisms in Erlang are designed for building fault tolerant systems as opposed to simply protecting systems from program exceptions.

Question 19

b. What is Chaos Monkey?

Answer 19

A software tool developed by Netflix engineers to ensure the robustness and resiliency of their Amazon Web Services by simulating failures of certain services bu shutting down some virtual machines.

Question 20

c. What is Formal Verification?

Answer 20

Mathematically proving the correctness of a design with respect to a mathematical formal specification.

Question 21

c. What is an advantage & disadvantage of Formal Verification? (2)

Answer 21

• Allows exhaustive testing of a system and defines an explicit understanding of the system
• Difficult and time consuming to write a specification using a mathematical proof instead of an English specification.

Question 22

c. What are two approaches to Formal Verification? (2)

Answer 22

• Model Checking

* Equivalence Checking

Question 23

c. What is Model Checking?

Answer 23

Verifies whether a particular set of properties holds true for a design.

Question 24

c. What is Equivalence Checking?

Answer 24

Compares two versions of a design to make sure they are functionally equivalent.

Question 25

d. What are Programming Language Innovations?

Answer 25

Improvements in programming language allowing software written in those languages to be more robust

Question 26

d. What is an advantage & disadvantage of Programming Language Innovations? (2)

Answer 26

• Allows more complex programs to be written in safety on knowing the software will maintain robustness
• Techniques such as using Virtual Machines, while are meant to improve robustness, also mean you have to be dependent on them so any faults in the Virtual machine will be a possible point of failure for a program.

Question 27

d. What are two approaches to Programming Language Innovations? (2)

Answer 27

• Virtual Machines

* Garbage Collection

Question 28

d. What is a Virtual Machine?

Answer 28

A separate operating system on a computer that simulates a separate computer, and in terms of programming languages this a program run on a specified virtual machine without having to be converted to machine code, allowing programs to be platform independent.

Question 29

d. What is Garbage Collection?

Answer 29

Automatic Memory Management, which recycles memory used by objects that aren’t going to be used again in programs.