Reading the “Register”:http://www.theregister.co.uk/ recently, I’ve been slightly surprised by not-so-sly digs at some Bloggers as a whole and specifically on their use of Trackbacks on a couple of their recent articles (see text in “Max OpenOffice Delay Debunked”:http://www.theregister.co.uk/content/4/32483.html and “Webloggers deal Harvard blog-bores a black eye”:http://www.theregister.co.uk/content/6/32315.html). Now as a long time register reader I’ve learnt to take their occasional rants with a heavy pinch of salt – indeed I’ve long suspected the register journalists are incapable of NOT having an opinion about something. In any case, their problem with Trackbacks (over and above the fact that they believe developers using them would be better served developing a port of OpenOffice for the Mac) was the problems it has been causing Google. Not really knowing too much about Trackbacks despite having them enabled myself, I decided to look into them in a bit more depth.
What are Trackbacks?
Probably the ‘bible’ definition of what Trackbacks are can be found over at the “movabletype”:http://www.movabletype.org website in a piece entitled “A Beginner’s Guide to TrackBack”:http://www.movabletype.org/trackback/beginners/. SixApart originally developed the Trackback system for use with MovableType, but since then the open specification has been adopted by several blogging systems. In a nutshell, Trackbacks are supposed to be a mechanism to inform a blog that something related to one of its entries has been posted elsewhere. For example, suppose I’ve made a post describing how I made a fancy image rollover using ActiveX which works in IE, and another blogger reads my post and works out how to do it in a way that can be viewed in multiple browsers. He would then send me a message letting me know this. This message is known as a Ping – this Ping causes my blog to record it and create a link associated with the original post.
So, why use them?
Theoretically Trackbacks can be used for a variety of things – it is after all a notification system. Matt Haughey over at A Whole Lotta Nothing is sending “Trackback pings from WinAmp(A Whole Lotta nothing – Trackbacks in Winamp)”:http://a.wholelottanothing.org/archives.blah/006625 to update his blog with the music he is currently listening to. That said most people use Trackbacks for the purpose of notification of related posts. In the example given above, the Trackback created by the ping not only lets the original author know that someone has read and commented on their original post, but lets readers of the original post find out more related information if they want to. This is why I use Trackbacks – I like to know when people may of had comments and expansions upon my posts.
And the problem is?
Well, there are several problems with Trackbacks – although many of these can be considered _potential_ problems
They break Googles Page Rank!
This is the Register’s “primary complaint(Google to fix blog noise problem)”:http://www.theregister.co.uk/content/6/30621.html. Googles page rank works by assigning a higher rank to a page which has been linked to ore times than some other page. Lets say I’m searching for a good article bob wrote. Along wth the original post, such a search can result in 100 AOL-style me-to blog posts (which I myself have been guilty of in the past) saying “look what Bob wrote, aint it cool?” before I actually get to Bob’s post. Now the link by me to the original post probably does want to feature in Google’s page rank- this is making the original post more important as people have found it interesting after all. The problem is that once I’ve made the trackback the original post then will update to show a link to my entry inflating my me-to style post’s Page Rank entry artificially, although my post contains little of relevance. Personally rather than the heavy handed approach proposed by Google of excluding blogs from their search results (which would have the affect of filtering some signal with all the noise) would be the ability to in some way mark a link so it doesn’t get taken into account by Google’s page ranking mechanism. Such a solution would of course need not only the cooperation of Google but would also require that all of the people implementing Trackback solutions to properly work.
Spammers could use them to ruin our Blogs!
The sending of a ping to a blog is a relatively simple affair. A bit of text and a link is sent to a blog, and the blog automatically assumes processes the link. In the case of my MovableType install, this results in the text and link being associated with my post. Whats important to note is that in no time during this process is any authentication of the server performing the ping carried out. It could be another blog, but equally it could be a spammer with a relatively simple client program trying to put ads on your blog. It could even be someone with a grudge against the blogger, using a ping to post defamatory content. Thankfully such things are VERY rare, however it may be in the future that some kind of moderation of pings will be required, or that a server should be authenticated in order to send such pings.
In his piece “Take Your Trackbacks and Dangle(Daring Fireball – Take Your Trackbacks and Dangle)”:http://daringfireball.net/2003/06/take_your_trackbacks_and_dangle.html, John Gruber believes that Trackbacks are not needed in the simple “this post relates to this other post” scenario:
And so while Six Apart calls TrackBack a notification protocol, the way people really use it is as a connection protocol. Person A sends a TrackBack ping to B to say that a post on As web site is related to a post on Bs web site. Assuming B does something with this TrackBack info, there is now a connection between A and B. But there already exists a mechanism for establishing connections between web sites: links. And there are ways to track links that are much simpler than TrackBack.
Referrers, for one. When you follow a link from one web page to another, your browser includes referrer information in the HTTP headers of the request. The referrer should be the URL of the page from which you came; if you click on any of the links in this article, for example, the web site youre heading to will get a referrer from this page at daringfireball.net.
John correctly points out some of the flaws to simply using Referrers – not least of which is the fact that a Trackback ping sends some text back to the original post giving some idea as to what the relationship between the two posts may be. This for me is a big stumbling block – John’s solution to the Trackback issue simply lists Referrers to his original post. If I’m reading his post why would I want to click on any of those links? They become simply a record for him, rather than something (potentially) of interest to his readers.
They require server side software to work!
Another gripe from “Take Your Trackbacks and Dangle(Daring Fireball – Take Your Trackbacks and Dangle)”:http://daringfireball.net/2003/06/take_your_trackbacks_and_dangle.html is the fact that the Ping and its handling requires software at both ends. Something has to _send_ the ping, and something at the other end has to _receive_ and process the ping. Blog systems such as Blogger and Radioland use a piece of software to produce a set of static files that do not require and special software to host (unlike MovableType for example). This immediately limits the potential users of Trackbacks and so reduces their usefulness.
Well, the perils of using Trackbacks are probably not limited to those posted above, but I think that Trackbacks are useful enough for me to carry on using them, at least until a better solution comes along. I do however think that the blogging community as a whole needs to look at the valid concerns raised by others and in part summarised here otherwise what in essence is a useful technology could prove in the long term more trouble than its worth.