Oracle Documentation: The broken links fiasco continues…

So I was just patting myself on the back for finishing my website clean up, then I happened on a few pages with broken links to Oracle documentation. That annoyed me, but I figured I better do a quick scan to see how many broken external links I had. The first attempt was a complete fail because the tool I used clicked all my Google Adsense adverts, making me a DotCom millionaire in about 3 minutes. I wrote to Google and apologised profusely. In my defense, the tool I used was right at the top of the list in the Chrome Web App Store…

Once I got a link checker that didn’t put me at risk of a jail sentence, things got a little more depressing. A very large number of my articles contain broken links to Oracle documentation. As I started looking at links it became apparent that Oracle have used at least 3 main URLs for documentation over the years:

• http://download-west.oracle.com/docs (8i -> 10gR2)
• http://download.oracle.com/docs (11gR1 -> 11gR2)
• http://docs.oracle.com/ (post 11gR2 stuff)

The versions listed are based on the links I’ve added in my articles. If you check today, all/most docs come from the “http://docs.oracle.com” address.

This in itself shouldn’t present a problem, because any company with an involvement in the web knows that URLs should never change. If by chance you do have to change something, you put a redirect in place. The problem is, Oracle don’t do this, or at least not consistently. Check out the following three URLs:

• http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96566/rcmdupdb.htm#441628
• http://download.oracle.com/docs/cd/B10501_01/server.920/a96566/rcmdupdb.htm#441628
• http://docs.oracle.com/cd/B10501_01/server.920/a96566/rcmdupdb.htm#441628

They are the same document, just using the three base URLs I mentioned previously. If you click them, you’ll notice the first one fails and the following two work. My guess is Oracle have created a 301 permanent redirect from http://download.oracle.com/docs to http://docs.oracle.com, but not bothered to maintain the http://download-west.oracle.com/docs URL, thereby breaking just about every link to its docs on the internet that references anything older than about 11gR1. That includes forums (including their own), blog posts, documents containing URLs etc. It’s just a nightmare.

So PLEASE Oracle:

• Stop changing URLs.
• When you do change them, PLEASE use rewrites/redirects properly.
• Remember, your rewrites/redirects should be permanent, not just long enough for search engines to update their indexes.

This would solve the vast majority of my gripes about the links to the Oracle docs…

Notes:

• For those not familiar with web servers, this kind of rewrite/redirect for a whole domain name is really simple. It’s one line in your “.htaccess” file, not a separate one for each page, so I’m not asking for the world here.
• I am aware there are other issues with changing URLs at Oracle that a blanket redirect would not solve. I’m not even going to start on whitepapers and PDFs…

Cheers

Tim…