How to start a C++ matrix class like this?

[nanchuangyeyu]

Hi,
I want to design a two dimensional matrix C++ class of myself. In prospect this class should implement such functions:
1. It should be able to be initialized by a binary data file(a .raw file without any header or format infomation.) with the file name as an argument of its constructor, if necessary, some side information such as the desired size of the matrix will be provided.
2. The following feature of ordinary matrix should be obtained:the summation of all the elements in the matrix.
3. The last requirement is this ordiary operation should be available in this matrix: the corresponding element product of two individual objects of this class(of the same size of course), i.e., the element of some position in the result matrix should be the product of the two elements of the same position in the two object matrices.

It's really a shame for me to list such detailed requirement but do not know how to start. For example, should the matrix be constucted in a traditional way or STL one such as vector. I am a newbie of C++ and just in a mess on this problem. Can some one give me some guide or is there some useful reference? Thank u in advance.

[PanGalactic]

From a design standpoint, the first requirement is a poor one. Why not have your matrix be initialized by an input iterator instead like STL containers? You can then pass a std::istream_iterator (or some custom stream iterator if needed) to the constructor. This way you haven't bound what could be a general-purpose matrix class to a specific file format.

A good starting point for software design are the SOLID principles: http://blog.objectmentor.com/articles/2009/02/12/getting-a-solid-start

Focusing on the Single Responsibility Principle will lead the way. You will want multiple classes working together to meet your requirements.

http://butunclebob.com/ArticleS.UncleBob.PrinciplesOfOod