Test Design Techniques Combining In Purpose To Build a Test Coverage

Test Design Techniques Combining In Purpose To Build a Test Coverage

There are some of the most well-known test design techniques: Equivalence Classes, Boundary Values, Pairwise, Decision Table, State-Transition Diagram.
You can use them separately, but it is better to combine them, for example, Equivalence Classes, Boundary Values, and Pairwise work well together. Let's talk about this.

Let's consider an explicit application with the creating events functionality in your own calendar as an example. This module has a simple form with fields to fill in and a "Create" button to create an event.

The fields are as follows: Name, Date, Time FROM, Time TO, Participant, Description, File.

The following requirements are formed for these fields:

  1. input field “Name”:
    - required (the field is required to be filled in)
    - unique (there cannot be 2 events with the same name in the database)
    - MAX characters accepted = 30, all characters allowed
  2. field “Date”:
    - required (the field is required to be filled in)
    - allows entering date manually in the format dd/mm/yyyy OR by selecting a required date from a date picker
    - only today’s date or future date is allowed
  3. fields “Time FROM” and “Time TO”:
    - required (the field is required to be filled in)
    - allows entering time manually in the format 00:00 (without seconds)
    - MIN interval allowed between times = 5 min
  4. input field “Participant”:
    - required (the field is required to be filled in)
    - only 1 participant can be entered
    - type = email (according to our imaginary logic of our imaginary application, the Participant will be sent an email with an invitation to join the event)
  5. text area “Description”:
    - NOT required (the field is optional to be filled in)
    - MAX characters accepted = 100, all characters allowed
  6. field “File”:
    - NOT required (the field is optional to be fill in)
    - only 1 file can be attached
    - allows opening the system folder from which we can select a file for attachment
    - MAX file size allowed = 5 MB
    - files of all formats and extensions are supported

So, having familiarized ourselves with the requirements for fields and form elements, we can form equivalent classes for each field. Let's start with positive equivalent classes (as an option):

Name:

  1. any valid unique value within 30 characters

Date:

  1. current date
  2. future date

Time FROM/TO:

  1. any interval within a day
  2. AM/PM transition
  3. 00:00-00:05
  4. 23:54-23:59

Participant:

  1. any valid email

Description:

  1. any valid value within 100 characters
  2. empty

File:

  1. any file of size 5 MB or less
  2. no file attached

Note that in some places (like in Name and Participant), we have set only one equivalent class, because it makes no sense to split into more (the logic of the application behaviour shall not change, regardless of whether we enter the event name in 10 characters or in 25 characters, or only in Latin, or only in special characters).

Of course, we can think of many more classes for each field, but the whole point of combining test design techniques is that we choose equivalent classes very carefully and only those that are more likely to help us find a bug. A larger number of "fictional" classes will only increase the number of unnecessary tests.

Now let's form negative equivalent classes (here the technique of boundary values will come in handy):

Name:

  1. more than 30 characters entered
  2. NOT unique
  3. empty

Date:

  1. past date
  2. not existing date (like 32/13/2002)
  3. empty

Time FROM/TO:

  1. time FROM is later than time TO
  2. not fully entered time (like 11:_ _)
  3. empty (both or one field only)

Participant:

  1. incorrectly formatted email
  2. empty

Description:

  1. more than 100 characters entered

File:

  1. file of more than 5 MB size attached

Now that we have a list of classes generated, we can generate test data sets that we will use for tests. To do this, we will use the Pairwise technique (to find unique pairs of values and reduce the number of unnecessary checks). A handy tool for this is PICT. Thus, we have the following:

POSITIVE CASES

  Name Date Time FROM/TO Participant Description File
1 any valid unique value within 30 characters current date AM/PM transition

any valid email

any valid value within 100 characters any file of size 5 MB or less
2 any valid unique value within 30 characters future date 23:54-23:59 any valid email empty any file of size 5 MB or less
3 any valid unique value within 30 characters future date any interval within a day any valid email any valid value within 100 characters no file attached
4 any valid unique value within 30 characters current date AM/PM transition any valid email empty no file attached
5 any valid unique value within 30 characters future date AM/PM transition any valid email any valid value within 100 characters no file attached
6 any valid unique value within 30 characters current date 00:00-00:05 any valid email empty any file of size 5 MB or less
7 any valid unique value within 30 characters current date any interval within a day any valid email empty any file of size 5 MB or less
8 any valid unique value within 30 characters future date 00:00-00:05 any valid email any valid value within 100 characters no file attached
9 any valid unique value within 30 characters current date 23:54-23:59 any valid email any valid value within 100 characters no file attached

 

NEGATIVE CASES

  Name Date Time FROM/TO Participant Description File
1 NOT unique past date time FROM is later than time TO

empty

more than 100 characters entered file of more than 5 MB size attached
2 more than 30 characters entered empty empty (both or one field only) incorrectly formatted email more than 100 characters entered file of more than 5 MB size attached
3 NOT unique empty not fully entered time (like 11:_ _) incorrectly formatted email more than 100 characters entered file of more than 5 MB size attached
4 empty not existing date (like 32/13/2002) not fully entered time (like 11:_ _) empty more than 100 characters entered file of more than 5 MB size attached
5 more than 30 characters entered not existing date (like 32/13/2002) time FROM is later than time TO incorrectly formatted email more than 100 characters entered file of more than 5 MB size attached
6 more than 30 characters entered past date not fully entered time (like 11:_ _) empty more than 100 characters entered file of more than 5 MB size attached
7 NOT unique not existing date (like 32/13/2002) empty (both or one field only) empty more than 100 characters entered file of more than 5 MB size attached
8 empty past date empty (both or one field only) incorrectly formatted email more than 100 characters entered file of more than 5 MB size attached
9 empty empty time FROM is later than time TO empty more than 100 characters entered file of more than 5 MB size attached

 

In fact, each of the lines (numbered 1-9) is our separate test to be performed. In this way, we do not test each field separately for positive/negative values, but the entire form as a whole, which allows us to get more reliable test coverage. Of course, you can generate more tests by entering all possible combinations of all equivalent classes into this table, but this is not necessary, besides, there will be too many tests. By finding only unique pairs with Pairwise, we have identified tests that are more likely to help us find the bug.

It's better not to combine positive and negative values in one test, it's better to separate positive and negative values so that you have a clear understanding which tests are of the highest priority.

Finally: All negative test data sets shall lead to the same result - the event shall not be stored in the system. An error shall be displayed next to all fields. However, with this approach, we may not immediately understand whether the event is not saved due to a negative value in one field or in several fields. Therefore, we can supplement the process of executing our negative tests:

  • once we have performed 1 test (1 row) of negatives, we change the negative value to positive in 1 field (any field) and click "Create" (the event shall not be saved anyway)
  • then again change the negative value to positive in another field and click "Create" again
  • and so on until we find the bug or come to valid values in all fields and to the successful creation of the event.

In this way we can find the following hypothetical bug: imagine that we have reached the point where we have one field with a negative value, all the others - with correct positive values. Clicking on "Create", we saw that the event was saved even though an error was displayed near the field with a negative value. This indicates that the functionality of displaying errors near the field works correctly, but the functionality of "preventing" the storage of events with negative values contains a bug that needs to be further investigated and entered into the bug tracking system.😉

Subscribed successfully

Subscribe to get our best content in your inbox

Chat