1. motivation

   1. we have a system

   2. this is a last look at the system

   3. first crack at the whole system, last crack at finding errors

   4. is the system good enough to be accepted by the client

      1. clients accept systems that do what they're supposed to do

   5. show the system works under certain circumstances - which circumstances

      1. the ones in which the client might be interested

   6. error, fault, failure

      1. error - an observable mismatch between system response and specification

      2. failure - the system behavior that resulted in an error

      3. fault - the part of the system from which a failure arose

      4. errors imply failures imply faults

      5. faults may not imply failures may not imply errors

         1. it is only in their observation that faults imply errors

            1. no observed errors does not imply no faults

            2. testing can show the presence of errors, never their absence

   7. can you test for correctness

      1. correct systems work under all circumstances - correct systems have no faults

      2. what are all circumstances - this is independent of client

      3. how big is the set of all circumstances

      4. how long will it take to test the set of all circumstances

      5. you can't test for correctness in general - omitting circumstances can lead to missing faults

   8. costly and difficult - budget 25% to 40% of total resources to testing

2. objectives

   1. testing as an activity itself, instead of an add-on

      1. testing as a managable activity

      2. make sure it gets done

      3. make sure it gets done will

   2. finding errors - testing

      1. tracing failures - debugging

      2. fixing faults - maintenance

   3. develop the widest, most relevent set of tests possible subject to resource constraints

      1. what the customer will accept defines the lower limit

This page last modified on 26 March 2001.