Linkbacks and Website Communication - part 2


Strictly speaking, Refback is not a linkback since it is not a protocol of communication between websites.

In other linkbacks, the linking site is responsible to inform the linked site about the linking document. But in this method, the linking site is not required to do anything and all the burden of finding the linking document fall on the linked site.

Refback is based on the fact that browsers used to add referer HTTP header to their document's requests. This referer HTTP header is added to the document request only when the user clicks on a link to the linked document in the linking document and its value is the URL of linking document.

In order to keep track with the linking documents, the linkback system should process each HTTP request and check whether it contains referrer HTTP header.

This linkback system has the some drawbacks.

First, the Refback method needs that someone will click on the link in the linking document. If nobody clicks on the link, the system will not be notified about the linking document

Moreover, due to privacy concerns, HTML 5 added noreferrer hyperlink annotation which indicates that the browser must not send referrer information when the user clicks on the link.

In other words, when following a link annotated with noreferrer, the browser must not include a referer HTTP header. Also, Some browsers allow the user to decide whether the referer HTTP header should be sent to the website when clicking on a link. (see Block Referer Headers in Firefox).

Another drawback, As my server log shows, referral spam become a trend these days. Any spam site that wants to advertise himself use this unethical technique. In this kind of spam, a bot sends a huge amount of document requests with referer HTTP header whose value is the spam site which is responsible to advertise and corrupt the site statistics.I find that unneglectable amount of HTTP requests with referer HTTP header are faked requests.

Since the Refback method relies on the goodwill of browsers to send valid and real requests, this method is not suitable for this task anymore.