4 ms·

I disagree. I think the accessor solution is the better solution in your example. You modification requires that File know about all extensions. It has to know

I disagree. I think the accessor solution is the better solution in your example. You modification requires that File know about all extensions. It has to know if something is an image, a document, a spreadsheet, etc. Every time there is a change, it needs to be updated. Thus you will have File object that is constantly updated. File version 2.2 won't process docx files (it thought documents ended with .doc). Also you place on File the burden of classification. is svg an image format or a text description. Also with your classification of File.IMAGE, File.Document, can you really generically process an image file. You might be able to, but it is hard. Whereas if you had the extension, with .jpg or .jpeg you load up libjpeg, .png you load of libpng, etc. With File.Document it is even worse. Can process document really generically process without knowing the extension? This is the reason I like accessors. Having them gives me flexibility in ways the original designer may not have foreseen.

Jare 14y ago
One thousand times this. The best way to design flexible and maintainable systems is to encourage small and simple objects and functions that can be asked to collaborate together via well defined interfaces. This collaboration will always require some objects to provide information that other objects and functions can use. In addition, the fact that part of an object's interface is the same as its implementation can mean that the implementation details are exposed (sounds bad), or that the implementation closely matches the interface (sounds efficient). Use your brain, not a canned rule, to decide which one of those is happening in your code.