Section 10.10. Why Exceptions (Now)?

10.10. Why Exceptions (Now)?

There is no doubt that errors will be around as long as software is around. The difference in today's fast-paced computing world is that our execution environments have changed, and so has our need to adapt error-handling to accurately reflect the operating context of the software that we develop. Modern-day applications generally run as self-contained graphical user interfaces (GUIs) or in a client/server architecture such as the Web.

The ability to handle errors at the application level has become even more important recently in that users are no longer the only ones directly running applications. As the Internet and online electronic commerce become more pervasive, Web servers will be the primary users of application software. This means that applications cannot just fail or crash outright anymore, because if they do, system errors translate to browser errors, and these in turn lead to frustrated users. Losing eyeballs means losing advertising revenue and potentially significant amounts of irrecoverable business.

If errors do occur, they are generally attributed to some invalid user input. The execution environment must be robust enough to handle the application-level error and be able to produce a user-level error message. This must translate to a "non-error" as far as the Web server is concerned because the application must complete successfully, even if all it does is return an error message to present to the user as a valid Hypertext Markup Language (HTML) Web page displaying the error.

If you are not familiar with what I am talking about, does a plain Web browser screen with the big black words saying, "Internal Server Error" sound familiar? How about a fatal error that brings up a pop-up that declares "Document contains no data"? As a user, do either of these phrases mean anything to you? No, of course not (unless you are an Internet software engineer), and to the average user, they are an endless source of confusion and frustration. These errors are a result of a failure in the execution of an application. The application either returns invalid Hypertext Transfer Protocol (HTTP) data or terminates fatally, resulting in the Web server throwing its hands up into the air, saying, "I give up!"

This type of faulty execution should not be allowed, if at all possible. As systems become more complex and involve more apprentice users, additional care should be taken to ensure a smooth user application experience. Even in the face of an error situation, an application should terminate successfully, as to not affect its execution environment in a catastrophic way. Python's exception handling promotes mature and correct programming.