In Apache, you can set up each directory on your server individually, giving them different properties or requirements for access. And while you can do this through normal Apache configuration, some hosts may wish to give users the ability to set up their own virtual server how they like. And so we have .htaccess files, a way to set Apache directives on a directory by directory basis without the need for direct server access, and without being able to affect other directories on the same server.
One up-side of this (amongst many) is that with a few short lines in an .htaccess file, you can tell your server that, for example, when a user asks for a page that doesn't exist, they are shown a customized error page instead of the bog-standard error page they've seen a million times before. If you visit http://www.addedbytes.com/random_made_up_address then you'll see this in action - instead of your browser's default error page, you see an error page sent by my server to you, telling you that the page you asked for doesn't exist.
This has a fair few uses. For example, my 404 (page not found) error page also sends me an email whenever somebody ends up there, telling me which page they were trying to find, and where they came from to find it - hopefully, this will help me to fix broken links without needing to trawl through mind-numbing error logs.
[Aside: If you set up your custom error page to email you whenever a page isn't found, remember that "/favicon.ico" requests failing doesn't mean that a page is missing. Internet Explorer 5 assumes everyone has a "favicon" and so asks the server for it. It's best to filter error messages about missing "/favicon.ico" files from your error logging, if you plan to do any.]
Setting up your htaccess file is a piece of cake. First things first, open notepad (or better yet, [url=http://www.editplus.com/]EditPlus2[/url]), and add the following to a new document:
ErrorDocument 404 /404.html
Next you need to save the file. You need to save it as ".htaccess". Not ".htaccess.txt", or "mysite.htaccess" - just ".htaccess". I know it sounds strange, but that is what these files are - just .htaccess files. Nothing else. Happy? If not, take a look at this [url=http://wsabstract.com/howto/htaccess.shtml].htaccess guide[/url], which also explains the naming convention of .htaccess in a little more depth. If you do use Notepad, you may need to rename the file after saving it, and you can do this before or after uploading the file to your server.
Now, create a page called 404.html, containing whatever you want a visitor to your site to see when they try to visit a page that doesn't exist. Now, upload both to your website, and type in a random, made-up address. You should, with any luck, see your custom error page instead of the traditional "Page Not Found" error message. If you do not see that, then there is a good chance your server does not support .htaccess, or it has been disabled. I suggest the next thing you do is check quickly with your server administrator that you are allowed to use .htaccess to serve custom error pages.
If all went well, and you are now viewing a custom 404 (page not found) error page, then you are well on your way to a complete set of error documents to match your web site. There are more errors out there, you know, not just missing pages. Of course, you can also use PHP, ASP or CFML pages as error documents - very useful for keeping track of errors.
You can customize these directives a great deal. For example, you can add directives for any of the status codes below, to show custom pages for any error the server may report. You can also, if you want, specify a full URL instead of a relative one. And if you are truly adventurous, you could even use pure HTML in the .htaccess file to be displayed in case of an error, as below. Note that if you want to use HTML, you must start the HTML with a quotation mark, however you should not put one at the other end of the HTML (you can include quotation marks within the HTML itself as normal).
ErrorDocument 404 "Ooops, that page was <b>not found</b>. Please try a different one or <a href="mailto:email@example.com">email the site owner</a> for assistance.
A server reponse code is a three digit number sent by a server to a user in response to a request for a web page or document. They tell the user whether the request can be completed, or if the server needs more information, or if the server cannot complete the request. Usually, these codes are sent 'silently' - so you never see them, as a user - however, there are some common ones that you may wish to set up error pages for, and they are listed below. Most people will only ever need to set up error pages for server codes 400, 401, 403, 404 and 500, and you would be wise to always have an error document for 404 errors at the very least.
It is also relatively important to ensure that any error page is over 512 bytes in size. Internet Explorer 5, when sent an error page of less than 512 bytes, will display its own default error document instead of your one. Feel free to use padding if this is an issue - personally, I'm not going to increase the size of a page because Internet Explorer 5 doesn't behave well.
In order to set up an error page for any other error codes, you simply add more lines to your .htaccess file. If you wanted to have error pages for the above five errors, your .htaccess file might look something like this:
ErrorDocument 400 /400.html
ErrorDocument 401 /401.html
ErrorDocument 403 /403.html
ErrorDocument 404 /404.html
ErrorDocument 500 /500.html
If you are interested, I have also put together a more thorough [url=http://www.addedbytes.com/apache/http-status-codes-explained/]explanation of the meaning of the HTTP Status Response Codes[/url].