I often have data that in some contexts I want to handle uniformly, but in other contexts I wish to differentiate it. For example, I could have a vector that is series of angles or a series of velocities, or a std::string that contains the current mode of a device or the name of that device instance. In some cases I want to operate on the specific information, such as modifying the angles or velocities using input data. In other cases I want to simply operate on all the data uniformly, such as serializing all of the strings or writing algorithms that will perform calculations for any device with a set of angles, or using a standard string algorithm. In yet more cases, physical devices will often have specific values that differ from one to another, like button differences between a wii remote and an xbox controller.
Has anyone put thought into designing and implementing data storage, algorithms, and device drivers in a maximally generic way and have insight or resources they can share?
One interesting document of note is the boost best practice handbook.
javac! (And from the tone and context of that mail, I get the impression he – the designer(!) – does not count himself among those three people.) – Jörg W Mittag Jun 30 at 19:56