Get it All
Together

It is always worth considering what changing your web layout will do to your hit rate.  If you’ve been out there a while, you’re going to need to consider that there’s some guy out there with a page of yours either bookmarked or even linked on his page.  When you delete that content and don’t follow up with a redirect, you suffer dead links and dead leads no one will bother to follow.  Jaimie Sirovich picks up with the 301 “Permanently Moved” linking discussion:

SEO Egghead by Jaimie Sirovich » Cashing In With Legacy Link Equity

Your computer sucks. The Core Duo is dead – now you need a Core 2 Duo! Well, at least nobody is selling them. And if the web page was located at /Core-Duo.html, one can 301 redirect /Core-Duo.html to /Core-2-Duo.html. This channels the legacy equity to related updated content. This practice is superior to returning a URL not found (404) error, both because it transfers equity contained by that URL instead of dispensing with it, as well as refers any old links to relevant updated content.

Over at my political site, DragonFlyEye.Net, creating the latest version of the site required many of the same considerations.  My last setup was largely produced with independent pages for the home page and each of the section home pages, with each calling a specific article ID from the database through a GET request (“index.php?art_id=25,” etc.).  When I went to the new “Clean URLs” schema, I had to provide a way that old links to content did not go away.

Even more vexing, the reason for moving to the “Clean URLs” schema was largely a security issue, so providing for the old style links meant allowing a certain amount of risk.  Moreover, making the changes via Apache was going to be inefficient because I go through a web host and would therefore need to rely on htaccess directives, which are much slower than httpd.config mods.

My solution, while not giving away the farm, was to use a fair amount of regular expression comparisons so that only those variables whose format I could predict would be allowed to work unmolested and all other queries would be pushed back to the home page of their respective sections. 

Now I am beginning to realize that my current naming convention leaves something to be desired, inasmuch as the URL does not include the title, which is a helpful SEO tactic.  Still, this is not the biggest problem the website has ever faced, and doubtless there are bigger fish to fry.  Perhaps in some later incarnation, this issue will be addressed; perhaps by the time some newer incarnation is developed, Google will have grown bored of title-bearing links.

Only time will tell.

Technorati Tags: , , ,

powered by performancing firefox