Producing Quality Software
Download as PPT, PDF1 like868 views
The document discusses the concept of software entropy, emphasizing that the quality of software declines as complexity increases and requirements grow. It introduces metrics for measuring software quality such as complexity, coupling, and conformity, and suggests practices for managing software maintenance to prolong application lifespan. Additionally, it advocates for controlling software quality to reduce costs and highlights the importance of using automated tools and code reviews to uphold standards.
Producing Quality Software
- 1. “ When a program is modified, its complexity will increase, provided that one does not actively work against this.” - Wikipedia “ What we need more of is Science” – MC Hawkins “ If you think good code is expensive, try bad code” - Anon
- 2. Software Entropy Entropy: Lack of order or predictability; gradual decline into disorder
- 3. Software Entropy Entropy: Lack of order or predictability; gradual decline into disorder Software Entropy: The decline of quality in an application as requirements increase
- 4. Broken Window Theory “ Consider a building with a few broken windows. If the windows are not repaired, the tendency is for vandals to break a few more windows. Eventually, they may even break into the building, and if it's unoccupied, perhaps become squatters or light fires inside.” “ Humans constantly monitor other people and their environment in order to determine what the correct norms for the given situation are. They also monitor others to make sure that the others act in an acceptable way ”
- 5. Broken Windows == Quality software? No. But given that Given that software entropy is inevitable People will produce code to the “common standard” We can conclude Poor quality software will degenerate quicker that better quality software Poor quality software will reach its end of life quicker Poor quality software is expensive Ignoring Complexity, Coupling & Conformity is expensive!
- 6. Controlling software quality controls software costs! $$
- 7. What are metrics? “ A way of applying a measurement to a quantity, which allow us to make a comparison” Used all the time in the real world “ Miles per gallon” “ Total cost of ownership” “ Balance of payments”
- 8. Software quality metrics Complexity Coupling Conformity
- 9. Software quality metrics - Complexity Measure complexity by Count the number of decisions (if’s) made along a logic path Count the number of separate modules (classes) invoked
- 10. Software quality metrics - Coupling Measure coupling by Count the number of possible data paths between tiers, modules and classes Count the number of public methods invoked between classes Double count and methods invoked against concrete implementations
- 11. Software quality metrics - Conformity Measure complexity by Number of deviations from a standard existing pattern
- 12. Measuring software quality Use automated tools when possible. Use written rules in code reviews. Use passing code reviews as a performance metric.
- 13. Using this software quality metric data Trigger code reviews when changsets exceed a threshold Extrapolate dates to when the application will need replacing at the current rate of change Manage the rate of change
- 14. Adding requirements without increasing software entropy Use the most general approach possible Factories Generics One size fits all If this is not possible ... Make the new code reusable Make your solution simple to understand! Refactor to make your change part of the whole.
- 15. Conclusion Every application has a natural lifespan. Maintenance should extend this lifespan, by fulfilling requirements at the lowest possible cost of software quality. Any major increase to complexity of the application should be subject to cost benefit analysis.