Basic oop Problem

Majid Kamali

Hi.
I am creating a Matrix class that represents a matrix and I have a problem with OOP concept.

which way is the best for representing matrix determinant ?
1. create a property that represents matrix determinant.
2. create a member function that returns matrix determinant.
3. create a library that performs matrix operations and within it, create a function that returns a matrix determinant.

Or for example Transpose that is not as common as determinant:
1. create a member function that returns matrix transpose.
2. create a library that performs matrix operations and within it, create a function that returns a matrix transpose.

this is just an example, but I have this problem in all my OOPs.

Thanks

3 Replies

Know the answer? Post it — somebody with the same question will find it here.

Vulpes15y ago
In my view, #1 is much better.

Exceptions should only be thrown if an 'exceptional event' occurs.

Square matrices are rarer than non-square ones so, in theory, you'd be throwing an exception more often than not if you use approach #2.

My personal preference is also to throw exceptions as little as possible to minimize the opportunities for the application to 'blow up' in the hands of the end-user.
Majid Kamali15y ago
I have a question from your answer.
I have two ways now:
1. derive a square matrix from a matrix base class and then create a read-only property named determinant.
2. create a generic matrix and a read-only property named determinant and if matrix rows and columns are not equal and then we call determinant property, it generates exception.

Which is better and why?
Thanks for your answers
Vulpes15y ago
If I remember correctly, only a square matrix can have a determinant.

So, I'd have thought that you should derive a SquareMatrix class from your parent Matrix class and then give it a read-only property to return its determinant. A property is better than a method as you don't need any arguments and you might want to cache the value in a private field to avoid having to recalculate it again.

With regard to the transpose, I'd just create a readonly property of the base Matrix class that returns the transpose. I'd use a property rather than a method for exactly the same reasons as for the determinant.