Question

Should the File class be used systematically when dealing with files or are there cases where it’s ok to use Strings to represent the path and filename?

Answer 1

If you take a look at the various file stream / reader / writer classes that take either a String or File, you’ll see that the constructor that takes a String is convenience overload for the constructor that takes a File. In other words, the String constructor typically creates a File from the pathname string and calls the other constructor.

So using the File version potentially saves constructing new File objects.

On the other hand, a File instance will occupy more memory than the equivalent String, and the cost of creating a File instance is small compared with the cost of opening and reading / writing a file.

The bottom line is that it probably doesn’t matter for a small-scale application from a performance perspective. With a large scale application the most appropriate approach depends on what the application is actually doing; e.g. whether or not it is repeatedly using the same files.

---

And as the comments say, there are other ways to “denote” files / paths as well; e.g. “file:” URLs and the Java 7 Path class.