Why are '/' allowed by web servers in the URL query string, when it really should be encoded as '%2F'?

I often see / (forward-slash) submitted in the query string of an URL, e.g. http://www.example.com?file=include/content.php - and the web server happily accepts and process the URL although the RFC standard dictates that / should be encoded as %2F.

Have I misunderstood the encoding? Are user agents free to choose either / or %2F in the query string, whereas / simply cannot go into the path segment of the URL? Are web servers not adhering to the standard, but instead trying to be as forgiving as possible?