That depends on how you use it. In our code (eg.
this file from last year), we make the DigitalInput part of the subsystem, then access it from outside with a more significant name (isForward() in this case). By doing it this way, the rest of the code does not have to care too much about the internal details, and if it turns out that the limit switch indicates the opposite of what you think, you can negate it in that named method, so it only has to be done in one place.