Evaluating Solutions in Programming

Study Notes

Solution evaluation is crucial after designing, implementing, testing, and debugging a solution.
Evaluations ensure the quality of the solution.
Evaluating involves asking basic questions about specific aspects of the solution.

Is it correct?: Does the solution truly solve the problem?
Is it efficient?: Does the solution use resources reasonably (e.g., time and space)?
Is it elegant?: Is the solution simple but effective? Is it well-designed?
Is it usable?: Does the solution provide a satisfactory user experience?

To be correct, a solution must solve the initial problem, regardless of speed or complexity.
Assume programs are incorrect until proven correct through thorough testing and debugging.
Correctness is vital in fields like science and law to avoid false positives.
Proving correctness can involve mathematical proof, but empirical testing is a more common approach in computer science.
Empirical testing involves carrying out a series of tests to validate the solution. Failure in any test indicates an incorrect solution.
Testing is an essential part of development. Problem specifications, and tests are prepared in advance.
The more tests that a solution passes, the stronger the evidence for its correctness.
A test plan example used vending machines

Algorithms require resources (time and space) to operate.
Efficiency measures how well an algorithm uses these resources, evaluated by considering time and space utilization.
Time: The duration of an algorithm's running time (from start to finish) measured as the number of steps.
Space: The amount of memory storage an algorithm needs.
Algorithm efficiency is assessed by analyzing time and space demands.
Different algorithms solve the same problems using varying efficiency. The 'best' algorithm depends on different factors
Complexity classes categorize algorithms based on their worst-case performance, indicating how resources would be used in the 'worst' situation for determining how efficiency is measured.
Table of common complexity classes provided (from O(1) to O(N!)

Elegant solutions are often challenging to create but deliver significant rewards in terms of both simplicity and effectiveness.
Elegance considers the depth of understanding of a problem in conjunction with innovative thinking for creating the solution, as simplicity is important
While different solutions can solve a problem effectively, their elegance might distinguish them.
Elegance isn't limited to software; it applies to disciplines like engineering, science, and mathematics. An elegant solution maximizes both effectiveness and simplicity.

A solution must provide a positive user experience.
Correctness, efficiency, and elegance are important but usability is paramount.
Solutions should be easy to learn, use, and forgiving of errors (user-friendly).
Usability measures how well a solution allows users to achieve their goals.

Learnability: How easy is it for a user to accomplish basic tasks when they first encounter the solution?
Efficiency: Once versed with the solution, how efficiently can basic tasks be performed?
Memorability: Can the solution be efficiently re-learned after some time without consistent use?
Errors: How many errors do users make and how easily can they recover from them?
Satisfaction: How pleasant or enjoyable is the solution for the user?

Solutions are seldom perfect because of factors like human errors or time constraints.
Optimizing one aspect can negatively impact another.
Trade-offs are also often inevitable when trying to optimize many aspects of a solution simultaneously.