My thoughts would be to look at the media type of requests. Typically you cannot get file names from the URL anyway.
This would likely be a text document http://www.mcafee.com/hello.txt
What about this? http://www.mcafee.com/something?paramx=123+paramz=abc
The second example is the more typical URL and doesn't provide any clear indication of the content returned. You could try following the link yourself, but many times these URLs would require session cookies, so you can't check them. Nevermind that checking each URL isn't very scalable.
In addition to that, nothing is stopping me from configuring my web server to treat ".txt" files as ".jsp" files. In that case, "hello.txt" could represent anything.
If you look at the media type, you get an understanding for the content of the data instead of file names which may not exist.