More Than One Answer More Than One Question

  • Slides: 9
Download presentation
More Than One Answer More Than One Question Written By: Michael Bolton Click here

More Than One Answer More Than One Question Written By: Michael Bolton Click here to view the article

When is 2+2=4 Wrong? Testing only a program’s correctness leaves out important issues of

When is 2+2=4 Wrong? Testing only a program’s correctness leaves out important issues of a program’s quality.

Testing from Multiple Perspectives n Quality Criteria ¡ ¡ n Structure Function Data etc.

Testing from Multiple Perspectives n Quality Criteria ¡ ¡ n Structure Function Data etc. Customer Criteria ¡ ¡ ¡ Capability Reliability Usability Scalability/Performance Installibility/Compatibility

Capability n n The traditional focus of testing Capability is a functional attribute Question:

Capability n n The traditional focus of testing Capability is a functional attribute Question: Can it do what is supposed to do? Does the program not do what is not supposed to do?

Reliability n Reliability is our ability to trust the program to perform the things

Reliability n Reliability is our ability to trust the program to perform the things of which it is capable n Reliability is also a security issue

Usability n Is it hard for other people or programs to use the program?

Usability n Is it hard for other people or programs to use the program? n 2 types: ¡ ¡ n Ease of use Ease of learning Often not considered because of its subjective nature

Performance/Scalability n Performance is not a question of “good” performance, rather if the program

Performance/Scalability n Performance is not a question of “good” performance, rather if the program performs as expected given its task n Scalability is a question of how it will perform will multiple users

Installibilty/Compatibility n How easily does the program install/uninstall? n Is it affected by other

Installibilty/Compatibility n How easily does the program install/uninstall? n Is it affected by other programs? Do other programs affect it? n

Conclusion By now we have all used programs that “worked” but were either too

Conclusion By now we have all used programs that “worked” but were either too unusable or too resource heavy too be useful. Good testing should account for the fact that customers – like us – do not like programs that are not user-friendly resource hogs.