Below please find three attitudes related to the most appropriate moment of starting planning for program testing.

Programmer A: I always start planning for program testing during program design or even earlier.
Programmer B: I always start planning for program testing while coding the program.
Programmer C: I always start planning for program testing as soon as the code is complete.

Please discuss the three attitudes and state your opinion about the right moment to start planning for program testing. Examples from your current or previous experience are welcome in the discussion.

Note. This discussion topic was adapted from textbook, Chapter 2, Exercise 32, page 149.

I'm sure that all three methods have their place, however limited each may be depending on the complexity of the program or system being designed and implemented.  I hope to be a programmer after I retire from the Air Force and predict that only one method is most effective for these large scale programs.

I think that programs that utilize programming teams to succeed would planning for program testing before the first line of code is written according to Programmer A.  Program testing allows programming teams to ensure that what they are about to write will meet the needs of the client, and ensure that most facets of the problem are identified prior to writing code.  The testing will lend itself to guiding programmers in the right direction before they even start.  Starting to answer the testing question right away would help to work out programming methods and a way ahead for the programmers.  I envision is as like having the answers to the test before walking into the classroom.

I have programmed Access databases in the past and due to the efficiency of the GUI interface and the use of forms, Access allows me to test as I go.  As long as I have my schema set up properly I can set up the necessary joins and design an efficient database.  Testing for me, using Access, is an ongoing process that starts with the completion of the first set of forms I generate.  Additionally, generating queries on the fly allows me to manipulate data quickly and with very little effort.  The change on the fly that queries allow, lets me test the code pretty much at the same time I am programming.  As long as I have a good set of data, which in my last database was already generated it was simple to test my programming while in the midst of programming.  When programming Access I do however think about what data I need extract and how to get there, so in essence I plan to test before writing any code, like that of Programmer B.

In the case of Programmer C, I think that failing to think about testing until the end of coding could be disastrous.  Both from the standpoint of money, but in coding time.  I have not had to write a large program with only coding, but the vast amounts of time involved with coding could all be wasted if the proper end result is not achieved.  That end result would be identified by the expected outcomes of testing and planning on testing.  This is method is like walking into the classroom without even showing up for class.  :)

Send me email

Visit my website