While perusing the inner-depths of the vast interweb late one evening, I ran across a specification in a protocol that fancied my interests to a great degree.
In other words, I ran across something cool.
Now I’m sure most of you nerds have heard of this specification before, but for some reason I never ran across this before. Apparently the rel=noreferrer tag or parameter was introduced a little while back in an effort for webmasters that do not want to pass referrer information to the website they are linking to.
In total there are dozens of these link relation tags out there. The most popular ones being “nofollow, author, etc”
I’ve read on several different sites that it is only supported by certain browsers, but I tested it on IE, Firefox and Chrome and all browsers seem to work in the sense that the logs don’t pass referrer information.
To test, I setup two hyperlinks that pointed to a server that I manage. The first hyperlink contains the “rel=noreferrer” parameter, the second one is just a plain old
<a href="http://elitedevel.com/">SOME Referrer</a>
<a href="http://elitedevel.com/" rel="noreferrer">NO Referrer</a>
As you might have guessed, when checking the log files the first link produces a log entry with all the expected information including referring IP and the page it came from.
76.109.XXX.XXX - - [28/Oct/2015:00:35:19 -0400] "GET / HTTP/1.1" 200 2932 "https://patrickcoombe.com/?p=2432&preview=true" "Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/46.0.2490.80 Safari/537.36"
Now with the “rel=noreferrer” parameter. Not only does it strip the referring domain, but the referring IP as well.
::1 - - [28/Oct/2015:00:35:06 -0400] "OPTIONS * HTTP/1.0" 200 125 "-" "Apache/2.4.7 (Ubuntu) PHP/5.5.9-1ubuntu4.13 OpenSSL/1.0.1f (internal dummy connection)"
Only thing that really crossed my mind is that if this tag really starts getting widely adopted, analytics companies are really going to be screwed.