In one of the interview I have been asked following question. Let’s say I have a class called ReportGenerator which has a simple function called GetReport.
This function have been used in CEO screen where it function will return reports of everybody in the company.
Now let say in future we have been asked to create two more screens. Manager UI and Employee UI. GetReport function will behave differently for this two screens. Like,
- Whenever Manager call that function he should get reports of all of his employees.
- When Employee call it, he/she will get only his/her report.
But when CEO calls this function, it will behave just like before. Means it will return everything.
- Add a new parameter called CallerType of type integer to exiting GetReport function.
- Add an If….else condition and return the correct report based on value of CallerType.
Problem with this solution –
In programming this scenarios is termed as scenario which violatesOCP (Open closed principle). Which says we should not change the existing software modules.
What happen if we violate OCP?
When GetReport function get changed there is high possibility of getting errors in CEO Screen (which was working before).
What will be the actual solution?
It was a design time decision. While designing the system we should have made GetReport function virtual. It make us override its functionality in new Derived class. After that we will just use derived class in Manager and Employee Screen
Below is a nice c# scenario based architecture interview question video.