It has been a while since I changed the backend of my site fron NewsPro to Movable Type. A side effect of this change was that all pages changed URL, which I considered a minor problem since most people enter the site through the home page (which did not change). Normally I try to keep URL's the same, but the old system used horrible filenames like 1025286960,58049,.html and stored all files in a single directory.
Like I said, most visitors enter the site through the fronpage, but there are some people looking for specific information with search engines. Until today these visitors were presented a '404 Page not found' error page of my provider directly followed by the homepage of my provider. In other words, these visitors end up at a completely useless page, and won't ever visit my site anymore!!!!
Something had to change. First, I wanted to change the error message from 404 (Not Found) to 410 (Gone), since the latter message is more precise. All visitors looking for never-existing files will still get the old 404 message. I read about this code a while ago on dive into mark, and I kept it in mind for using it on m own site.
Secondly, I want to show them another page, which at least links to my site. For the moment I chose to present them the search page, so at least they can search my site for the information they are looking for. I am planning to create a separate error page explaing what has happened in more detail, but until now I will just show the search page.
So how did I set this all up?
The Apache server has the ability to specify configuration options in a file called .htaccess. The settings in this file are valid for the current and all underlying directories, and thus can be overruled by another file. I created this file in the root directory of my site and added the following information:
ErrorDocument 410 /cgi-bin/mt/mt-search.cgi
RewriteEngine on
RewriteBase /
RewriteCond %{REQUEST_URI} ^/cgi-bin/newspro
RewriteRule cgi-bin/newspro/(.*) - [G,L]
In 'normal' language, these lines say the following:
The first line says that the page at /cgi-bin/mt/mt-search.cgi (which is my search page) has to be shown in case a request has status 410: Gone.
The second and third lines activate and configure a module for the Apache webserver calles mod_rewrite, which allows me to process the next two lines.
The fourth line defines the condition to me used for the rewrite. It says: if the requested URL starts with '/cgi-bin/newspro'.
The last line specifies what has to happen if the condition above is met: rewrite 'cgi-bin/newspro/*.*' with nothing (-). Besides the rewrite (which does nothing) two options have been specified: G and L. The first changes the status to Gone (410), and the L tells that this is the last rule to process.
In short: if the URL contains /cgi-bin/newspro, the status will be changed to Gone, and when error 410 (Gone) occurs, the search page will be shown.
; "Twitter conversation")
; "Movable Type")
; "CeBIT")
; "ClamWin updating")
; "Google Notebook - New!")
; "Events report by file type")
; "Google Analytics")
; "Jesus at Christmas")
; "from e-mail to Enterprise Social Software")
; "Eleven")
Surely if these pages still exist, but the URI has changed, you should issue a 301 (Moved Permanently) response pointing to the new URI?
410 is only if the pages have been deleted.