Stop the trackback myth

The heading of a blog home page usually contains some statistics like the number of posts in the blog and the number of comments. Almost always it will mention the number of trackbacks. The codebetter blog homepage has a list of bloggers (reverse) ordered by number of posts, with our trackbacks. The way the latter changes over time made me suspicious. There are liars, big liars and (blog) statistics so I would like to know what I am actually reading.

A trackback is a word everybody is using, but almost nobody can give a good definition what it really is. I went to wikipedia :

From Wikipedia, the free encyclopedia.

TrackBack is a system implemented by Movable Type and later adopted by many blogging tools, including TypePad, WordPress, and Nucleus, that allows a blogger to see who has seen the original post and has written another entry concerning it. The system works by sending a ‘ping‘ between the blogs, and therefore providing the alert.

TrackBack typically appears below a blog entry and shows a summary of what has been written on the target blog, together with a URL and the name of the blog.

The specification for TrackBack and its implementation were developed by Six Apart, creators of the MovableType weblog and content management system.

Aha, so it is invented by a specific blogging tool. And not the most popular one amongst developers. What I know about the way .text and cs implement trackbacks when posting is that they scan a post for links and will send a trackback to all links found. And it’s up to the receiver what to do with them. What I have noticed about the way .text and cs deal with incoming trackbacks is that it’s next to random. The most extreme examples I’ve seen is referring to some of my own posts and seeing my count rise by six. And there are lots of links to my post which didn’t influence my count at all. So the tools used to post and to refer determine if a trackback is seen in the statistics.

What’s of real interest are referrals. In the statitstics of your blog it’s a list of links which have been followed to find the post. Including links from google where you can see what search phrases have been used to find your story. Referral is also in wikipedia but no-one has taken the effort yet to enter it as an internet term.

To sum it up:

  • Trackbacks are no open standard
  • Trackbacks depend on the blogging tools used
  • Referrals tell the real story

Should CodeBetter switch from displaying trackback count to referral count ? Or to page hits ? A side problem would be we’de have to swich to scientific notation :) Imho we should display nothing, it’s the content itself which counts.

 

This entry was posted in Chatter. Bookmark the permalink. Follow any comments here with the RSS feed for this post.
  • Pingback: traffic travis review

  • http://codebetter.com/members/pvanooijen/default.aspx pvanooijen

    And now I’m going to close the possibility to comment on this one. Due to trackback spam :)

  • http://codebetter.com/blogs/peter.van.ooijen/ pvanooijen

    Frans:

    OK. misuse, that’s a another (potential) drawback. Point taken

    Scott:

    I’m sorry you don’t like cs or .text, but a lot of .net specific blogs are hosted by them. And as long as they don’t implement trackbacks the way sixapart and the likes wants it trackbacks are no standard. Let alone an open standard. The google result you mention is by one of those Wikipedia mentioned companies/institutions.

    The way you describe a referral is exactly how cs keeps track of referrals. An extra property is indeed the number of hits. So that’s perfect.

    Your vote does count, you are reading this blog, and take the effort to comment. What do you take us for ?

  • http://www.lazycoder.com Scott

    So the rest of us that aren’t using the craptacular CS or .Text blog engine use trackbacks in a consistant manner.

    “Aha, so it is invented by a specific blogging tool”

    As were RSS and RDF, so?

    “And not the most popular one amongst developers.”

    What developers? Just you guys? ASP.NET developers? Perl developers? Java developers? C++ Developers?

    “Referrals tell the real story”

    Nope, a trackback is a specific event triggered intentionally and should be counted one time. A referral may not be. So if I click on a link from my blog to this post 50 times, should your referral count go up by 50 or just 1? If you group unique referrals together and count them, well you essentially are doing the same thing trackback is intended to do.

    “Trackbacks are no open standard”

    errr, I guess it depends on what you mean by “open standard”. I type “Trackback” into google and voilla, the first hit is the explanation of how trackback should work. There’s even a link in there explaining the REST mechanism. http://www.sixapart.com/pronet/docs/trackback_spec

    Seems pretty open to me. Don’t blame trackback for CS’s failings, you wouldn’t like trackback when it’s angry. 😉

    My vote would be to not show the stats for the reason you state above, however my vote doesn’t count since I don’t blog here.

  • http://weblogs.asp.net/fbouma Frans Bouma

    … and, trackbacks are now used to spam blog comments!

    Chinese spammers now use tools to add a trackback to a blog which ends up not as a link to a blog but to a website which the spammers wants to draw attention to (or get ranked higher in google).