Better Trackback?

There is talk of pushing for Trackback to become a standard. A few of the problems with Trackback are immediately apparent: horrible internationization support, bad auto-discovery, proclivity for spamming, no verification, historical baggae of category junk, bad spec. Fix all these and you get… pingback. Pingback is big enough now to make a blip in Google’s markup survey, and is supported by a wide range of platforms. The question is whether people are going to want to support an existing and robust standard or want to put their name on something new, the global “not invented here” syndrome where everyone wants their 15 standards of fame. (As someone who has been involved in several standards myself, I admit the draw is strong.) What Pingback does need is a better advocacy site, like atom has.

17 replies on “Better Trackback?”

I’ve never been a big fan of either Trackback or Pingback, just because it duplicates an already good mechanism for notification (the HTTP referrer header). True, there are advantages to pinging (built-in to blog software and potential for easy post excerpts), but I wonder if near real-time tools like Measure Map and Mint will change that. What’s missing is a software package that does something smart with the referrer data for re-publishing.

Manton, the reason referrer doesn’t work as well is that (a) it’s often and easily faked and (b) the top referrers usually aren’t the permanent link for a resource. You might get 1000 hits from, but what you really wanted was the permalink to the entry he linked you from. That’s what Pingback does, it works like a referrer but with verification and permalinks.

Manton: Trackback and referrers tracking don’t really have that much in common. TB is a lot more than a notification gizmo… it’s essentially a means of posting remote comments.

Where a referrer is saying “someone followed a link from A to B”, Trackback says “someone at B thinks that the readers of A will be interested in what B has to say about this”.

I’ve never liked Pingback. The summaries that it send can be non-sensical – see for an example. With Trackback, the summary sent is usually the first few words in a post and may even be properly customised.

Still, my suggestion would be those involved in the design of the Pingback standard should get involved with standardising Trackback. It may well be that Pingback, or a modified version thereof, becomes the accepted Trackback standard if the arguments in favour of it are good enough. It is, after all, widely deployed as part of WordPress and other blogging packages.

But as for Trackback, it needs improving and so I’m glad to see this process being commenced.

The excerpt is at the discretion of the receiver. The code in WP that extracts the excerpt isn’t as good as it could be yet, there are also plugins to make it better.

I assume I’m still the only person on the planet who thinks that both *backs are good things, for different purposes, and should both be embraced? (For those I haven’t bored into a stupor over the years, Pingback says “here’s a URL that links to you, do as you will” and Trackback says “I think that someone who read you would be interested in reading me, here’s a link and a teaser to bring them in.”)

Shame, that, I’d really like to see a combined I-D for both.

If I thougt ID was the answer to spam, I might be inclined to that as well. (It’s not.)

I hardly get trackbacks anymore, but I’ve never approved an asynchronous trackback, it just seems tacky.

Matt, thanks for the extra insight. I understand the reasons, but would still argue that better use of referrers can accomplish nearly the same thing. If the incoming URL is a blog, grab its RSS and find the post that linked to you, using that for verification and as the basis for excerpts and permalinks. I admit that has its own problems (more network intensive, complex, and prone to failure), but I like that it places the burden on the receiving end instead of reinventing a fundamental piece of the web.

And Roger, the distinction you make between linking and remote commenting is interesting. Seems a subtle difference, but good point.

Matt, I think the Internet community would benefit a great deal by you joining this effort and sharing with the working group what you have learned in your own efforts to improve upon the protocol. In trying to address these problems myself, I chose TrackBack as a starting point only. What we end up with is likely to be different.

We need to build concensus around something that is fundamentally more useful and secure for everyone. Also, I am hoping that by giving this process over to a standards body that we will do our best to divorce ourselves from the pride and prejudice of ownership. If this is to succeed, it must be community effort, not one by Six Apart alone. Wouldn’t you agree?

Also – I am hoping that the first cut at a revised spec ( you will see that it attempts to address a few of your concerns right out of the starting gate: a) hopefully the spec is not as “bad” as it was, and b) autodiscovery is easier. But I would love to hear your thoughts directly. Although those issues are the “easy ones.” 🙂

Matt, I have subscribed you manually. Trust me when I say how embarassed I am about my need to do this. I took it for granted that everything would work flawlessly when I setup the list. I had no idea that we would have some script that would wake up every 15 minutes and lock down any public mailing list.

But I know and trust it will all get ironed out soon, and Mark “Junior” Smith, the creator of LifeWiki and LiveJournal developer extraordinaire just fixed the attachment bug. So slowly but surely everything will be as it SHOULD be. My sincerest apologies to all those who have experienced the slightest difficulty thus far.