June 14, 2003

Trackback hype debunked

Joshua picks up on the trackback conversation, but there are many problems with the comparisons of trackbacks and referrer tracking (I am posting this for clarity not to poke at my friend Joshua and comments are not turned on for his entry or I would post there). It sounds like referrer logs are not set properly in what Joshua is using for a comparison to trackback. It may be that Movable Type is not set to take full advantage of referrers, which is sad. Joshua and the posts he links to explain trackbacks properly.

Of the three comparisons of trackbacks to referrer only one is correct, trackbacks do not require an actual hyperlink to the article to work properly (why one would not have a link to what they are discussing is odd as it is the Web). The other two comments are not correct if referrer logs are setup properly. In my last 100 referrers page I can see exactly what page a link came from (I used Charles Johnson's referrer as a base for mine). If the hyperlink is clicked from a permalink page I can see the exact page the link came from and if it links to my permalinked entry I can have categorical sorting also.

I have not seen a need for trackback on my site my referrer log (real-time) and access log show me exactly what is going on. Oddly there is always a lot of talk about trackback on blogs, but very rarely are they ever used. This may point to a usability problem with trackback. Referrer and access logs on the other hand do not need this extra step for the site owner to get the information.

Once I get unburried with paper work for work and have a clear head that the completion will bring I may implement a referrer version of trackback for public consumption, unless the lazy web does it first.

Posted Comments

Hi Thomas. Thanks for the comments.

"If the hyperlink is clicked from a permalink page I can see the exact page the link came from"

Yes, but unlike TrackBacks, referrers don't create any bound relationships between links and entries, they only create relationships between pages. This is what I meant when I said, "TrackBacks create links between individual permalinked posts, whereas referrers only create links between websites." I probably should have said pages instead of websites.

"if it links to my permalinked entry I can have categorical sorting also"

Yes, but I'm guessing you would have to write a script to sort the categories. TrackBack, on the other hand, has the built-in ability to sort categorically because of the entry to entry (or in this case entry to category) relationship. No extra work is required. That is what I meant when I said "TrackBacks allow category-specific aggregation, whereas referrers canĂ­t aggregate anything."

"why one would not have a link to what they are discussing is odd as it is the Web"

It's odd if you don't consider TrackBack. I'll give two examples.

  1. Say I wrote an entry about new research in flea prevention for dogs. In my initial entry, I didn't link to anything because I didn't know of any web sources that were close to what I was discussing. A day later, I realize that someone else wrote an extensive entry on the same subject a week earlier. I could send a TrackBack ping that would create an entry to entry relationship, and I wouldn't ever need to link to the other entry from my website.
  2. I can write an entry about weblogging, but again not link to anything. At the same time, I can send a TrackBack ping to the Blogpopuli TrackBack URL (http://www.blogroots.com/tb_populi.blog?id=1) for the category specific aggregation I mentioned above.

"comments are not turned on for his entry or I would post there"

Finally, I should say that if you were to use TrackBack, you could send a ping to my entry, and a link to your comments would appear on my site. If you don't use TrackBack, you can use Adam Kalsey's new SimpleTracks to link your entry to mine.


Joshua, if the creator of content clicks the connecting link from their permalink page to my permalink page it does track exactly as trackback. I have many of these in my access logs, which are relatively easy to parse. I use this on a regular basis to find who is linking to what and often from exactly where.

I use my referrer logs to then go check my access logs and do a quick search to find the same info as trackback. This is an easy script form me to write. But, I do not use MT (I really like MT, but I also like having my own tool to give better understanding of what is actually going on) which means I do not have somebody elses script to use. Trackback is not magic it is just prescripted.

Ah, I can aggregate links to permalinked entried very easily and MT could too. All permalinks are connected to their corresponding categories. I check the access logs every morning to find other information related to what I posted based on what others have linked from.

The magic of trackback is in having permalinks, which access logs and properly built referrer logs can easily use. By linking to your permalinked page you should have an entry in your referrers and access log that will give you identical information that trackback does.

In my referrer logs and access logs I see link from Tantek's "end of Mac IE" posting to my permalinked post.

Lastly the "comments" should be labelled "trackback" for clarity (if that is what they are) as the two are very different constructs.

I continually consider adding trackback, but have not found a great need as of yet. Some of what you have mentioned has helped me see some benefit, but I have not made it over the top yet. Some of what others use trackback for, adding correlations between content, I do in my comments after the fact. I know Google and other search engines use this correlation and having the hyperlink on both specific pages reinforces the relationship between the content.

All the best and thanks. (I really need to note that HTML works in this text area).


Comments are closed.

This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License.