|Developer Diary · You Heard It Here First · Thursday 26 February 2004|
|Principles of Programming: Input Validation|
One of the simplest principles of programming is to validate input yet this chore is the single biggest failing of all computer programs. Most of the virus vulnerabilities that exist as well as most bugs are caused by programs receiving unexpected input. I can understand why most programmers do not validate internally generated arguments to function calls but it is incomprehensible to me that so many programmers do not validate input coming from outside the program yet it seems every program out there fails in this regard.
Some people have made a pastime out of crashing well-known programs this way. I remember back around 1996 when NT was new there was a guy who wrote a program that could systematically call every function in its API. He would just feed the API as many possible inputs as he could and then log all the crashes and blue screens. I looked at his log (which was being sent regularly to Microsoft). There were literally thousands of crash combinations listed. Nowadays MS has wised up a little so there are fewer problems, but nevertheless new buffer overruns and other serious input validation failures come up every month even in major operating systems like Windows.
The principle of validated input is actually composed of four seperate principles:
Assume that input may be missing
Assume that input may be out of range
Assume that input may be malformed
Provide for a way to reject input
These are the most basic of programming principles. Taking the time to follow them will improve the quality of your programs to an extent greater than any other single measure you can do.
|return to John Chamberlain's home · diary index|
|Developer Diary · about · email@example.com · bio · Revised 26 February 2004 · Pure Content|