Minimal test cases to catch bugs in the Calendar Date to Day convertor

 

Here’re the fields and respective input values against which the Date to Day convertor can be tested (left-side-table):

Month	Day	Year				Month	Day	Year
1	1	2010				2	28	2010
2	2	2011				4	29	2011
3	3	2012				12	30	2012
4	4						31
5	5
6	6
7	7
8	8
9	9
10	10
11	11
12	12
	13
	14
	15
	16
	17
	18
	19
	20
	21
	22
	23
	24
	25
	26
	27
	28
	29
	30
	31

However, to break the software, a smaller set of values is sufficient enough!

We can apply Boundary Value Analysis method at the upper end of days range to consider:

1. 1 case of a month having max 30 days
2. 1 case of a month having max 31 days
3. 2 cases for the month of February (max 28, 29 days)

Thus, our set of values reduces to the second table shown above. Here we have randomly picked months 4, 12 and 2 respectively.

Let’s take a closer look!

To consider both positive and negative test scenarios, we consider:

1. An even numbered year (2010)
2. An odd numbered year(2011)
3. A leap year(2012)

Now even this reduced set of values for Year, Month and Days of Month can lead to an overwhelming set when we consider the possible combinations that can be used to input data in the form. The number would be

= (Number of Month Values) * (Number of Year Values) * (Number of Day Value)

= 3 * 3 * 4

= 36

However, there’s a more efficient scientific technique available!

So, let’s feed the reduced set of data to the All Pairs program from satisfice.com

Here’s the raw set of test cases derived by the program:

Month	Day	Year
2	28	2010
4	28	2011
12	28	2012
2	29	2011
4	29	2010
12	29	2010
2	30	2012
4	30	2010
12	30	2011
4	31	2012
2	31	2010
12	31	2011
~2	29	2012

Categorize it:

Test generation method	Month	Day	Year	Test type
Program generated	2	28	2010	+ive
Program generated	4	28	2011	not required
Program generated	12	28	2012	not required
Program generated	2	29	2011	-ve
Program generated	4	29	2010	not required
Program generated	12	29	2010	not required
Program generated	2	30	2012	-ve
Program generated	4	30	2010	+ve
Program generated	12	30	2011	not required
Program generated	4	31	2012	-ve
Program generated	2	31	2010	-ve
Program generated	12	31	2011	+ve
Program generated	~2	29	2012	consider as additional case for Month value 2 only
Additional cases	2	29	2012	+ve

Filter it:

Test generation method	Month	Day	Year	Test type
Program generated	2	28	2010	+ive
Program generated	2	29	2011	-ve
Program generated	2	30	2012	-ve
Program generated	4	30	2010	+ve
Program generated	4	31	2012	-ve
Program generated	2	31	2010	-ve
Program generated	12	31	2011	+ve
Additional cases	2	29	2012	+ve

Execute these and report the bugs!