Considerations for Page Not Found Errors

By Isaac Ferreira
Posted Tuesday, September 7, 2004

Abstract: Good error messages are polite and informative. Polite means using simple language that does not imply the error is the users fault. Informative means suggesting steps on the error message page the user can take to correct the problem.

It is inevitable that mistakes will be made on a website. Although no user wants their work interrupted with a page not found error, it does provide the website an opportunity to demonstrate the level of care taken to ensure a good user experience. A 404 Page Not Found error can be a rather helpless experience. This is why we recommend some of the following actions:

Correcting mistakes automatically
Avoiding difficult file names
Explaining why links disappear
Offering alternatives for finding the correct page
Forcing Internet Explorer 5+ to display your 404 and not it's own

Correcting mistakes automatically
The friendliest thing a website can do in the situation where a page is not found is recognize the error and correct it so a user never knows.

For this to take place problem URL's must be tracked to develop an understanding of what pages are being reached in error most often. That discussion is better left for another article, but some common mistakes such as file name misspellings can be fixed and redirected before the 404 error page is ever needed.

Although it may seem a black art, it's not. Common problems like typing an extra "l" at the end of an "html" file name are easily spotted and quite common. Also, common file name misspellings can be found by looking through error logs and spotting them.

For situations which are not so obvious, software can be written to scan parts of a website and give the user potential alternatives to the correct webpage. Although this does not "autocorrect" per-se, it does narrow the possibilities which will always help.

There are situations that little can be done. Trapping the referring page (Such as using the "HTTP_REFERER" server variable) and automatically sending an email message to the Webmaster of the site is a basic step that can be taken and generally brings good results.

A more complex situation such as typing "(www.mysite.xom)" is not something that a webmaster can manage. No contact can ever be made with the web server because the extension is incorrect. Solving this would require cooperation from large organizations but we feel it is worth mentioning to promote the idea.

Avoiding difficult file names
Long and complicated file names are a common reason that 404 errors are generated. Simple file names are easier to remember and easier to type. Unfortunately many content administration tools auto generate file names such as "2002ers33455_43549042.html" which are mostly useful as a copy and pasted link in an email or hyperlink on a website, but not much else. Not many people are going to remember a name like that. Some websites, such as news outlets, do not have a lot of choice because they create upwards of hundreds of new web pages a day. So avoiding difficult file names is not always possible but suggested.

Explaining why links disappear
The most common reasons for the 404 error message are:
The page was moved elsewhere on the site
The website was redesigned and connections to older content was not made
Another site linked to a page and got the link wrong
The link was dated or the page simply no longer exists
A search engine has an old index and hasn't updated
The user typed the address incorrectly
We recommend listing these on the 404 error page to help educate users. A significant percentage of websites do no have error messages and here users can pickup a few tips that can be useful in the future.

Offering alternatives to find the correct page
Offering alternatives to help the user find the correct information is the single most important thing an error page can do. Before we list suggestions, don't redirect an error automatically to the home page. This is disorienting. We recommend the following be included on a website error page: If you cannot script "targeted" alternatives, add a listing of the most common mistaken URLs on the website (studying log files highlights what errors are made the most) Offer a search field linked to the site's search engine: even when users can't guess the current URL, they may be able to specify a query to locate the desired information. Offer a link to the site map: this will help those users that might have been just looking for a particular section, say, the "Services" page Offer a link to an "About the site" page, if there is one Link to a contact form that the user can fill in to ask about what they were looking for Forcing Internet Explorer 5+ to display your 404 and not its own Version 5+ of Internet Explorer has its own 'friendly' error messages built in. Some people who have designed their own custom 404s have found that IE ignores their page and displaying its own message. To get around this make sure that the size of your file is greater than 512 bytes. If you just add a graphic, you'll be well over that size, this way Internet Explorer choses your custom error instead of its own.

Article by Isaac Ferreira, visit (www.onlinetransmissions.com) for related articles.

About the Author
None