TL;DR
This blog post is about important attribute that almost every data has, unique id.
I have several ids:
- Croatia government OIB, unique in Croatian context
- Facebook id, unique in Facebook context,
- …
In SFDPOT mnemonic [Bach], D is data. So what test ideas can we have thinking about data?
Every data must be uniquely identifiable. So we have data ids. Sometimes several ids. There is you test idea:
Do we use proper id in this part of application?
For example, user has id_a, id_b, id_c. Application sends id_b to external system.
What is id_b?
Maybe we should send id_a to external system?
Look at other application functions that communicate with same external system, what do they send as user id?
Application that you are testing, has its own ids. Do not stop your questioning when you confirm that id_b is successfully sent to external system.
Testing is not about showing that the product can provide an expected result. Testing is exploring and experimenting to discover and learn about the unexpected [Bolton].
Question the data, do we use proper data in this function?
