LawMeme LawMeme Yale Law School  
LawMeme
Search LawMeme [ Advanced Search ]
 
 
 
 
Features: Trackback and Free Speech
Posted by James Grimmelmann on Wednesday, June 18 @ 04:03:40 EDT Free Expression
I love Trackback. (Granted, I've never used it myself, but still I love it.) Trackback has the potential to be the most important thing to happen to the web in years- because it's one of the few recent developments with the genuine potential to redefine the way we think of the web.

In particular, I think Trackback and technologies in the Trackback vein can help us look at certain online free speech issues in a new and more optimistic light. Full details inside . . .

Technical Background

(If you're already familiar with Trackback, you can skip ahead to the next section.)

First, some background. Trackback is an innovation from the good folks at Six Apart, the makers of Moveable Type. Moveable Type is a piece of blogging software, now available in both hosted and roll-your-own-hosting versions. It distinguishes itself from the zillion other blog tools out there mostly in the clarity of its interface and its well-articulated API for third-party extensions. Trackback made its appearance on the world stage as a feature rolled out as part of Moveable Type. Yours free with every download; just check a box on your options page to enable.

To understand how Trackback works, it's helpful to start with blog comments. When you post a comment to a blog, you're sending a message to the server the blog is on. That message contains a bunch of information, including both the text of your comment and some metadata about it, such as your name or the URL of your home page (depending on the blog software and which information you choose to supply). It also includes enough information to identify the specific blog entry on which you're commenting. The server then modifies the HTML of the blog entry to indicate that you've left a comment (some blogs just show the comment immediately; others ask readers to click through to a comments page).

Trackback does exactly the same thing, except that the message to the server (confusingly called a "ping") doesn't contain the text of the comment. Instead, you send the URL to the page where the text of your comment can be found. The server doesn't (generally) go and download the text of the comment. Instead, it modifies the HTML of the blog entry to indicate that you've left a comment, typically by providing a hyperlink to the URL of your page with the comment.

The genius part of Trackback's design was the realization that the obvious example of another page with a comment is an entry on another blog. Moveable Type integrated Trackback into its interface, such that if you write a blog entry which comments on my blog entry, you can send the "ping" to my server at the same time as you post the entry to your blog. This means that readers of my blog now know that you've written something following up on my entry and can immediately click through to read it.

Officially, Trackback is a protocol, rather than a Moveable Type feature. Other blog programs are adding Trackback functionality, both to send and receive pings. It's also a protocol whose semantics are deliberately open. Right now, the usual convention is that Trackbacks are displayed as a list of links, but there's no ironclad rule saying it has to work that way: you could just as well suck down the HTML from the other end of the link and display it inline (and if the URL was a URL to an XML file, rather than raw HTML, this could work out quite elegantly). Or, experimenters with Trackback could use it to send all sorts of helpful metadata (e.g., of the sort usually now shoved into an RSS feed) about the nature of the link. The point is that Trackback defines a floor, rather than a ceiling.

Trackback and the Two-Way Web

What I love about Trackback is that it turns links from being one-way creatures into being truly bidirectional. Think about it. The traditional hyperlink is the equivalent of someone standing on a mountain pointing a telescope at another mountain miles away. If you look through the telescope, you can see enough of the other mountain that you could go over to it and find the relevant spot. But you as a hiker, standing on a mountain, have no way of knowing whether there are telescopes pointed at the spot where you stand. You can't easily find out what that spot looks like from other perspectives.

What this means is that you can only really read conversations on the Web the same way you read blogs: by going backwards in time. You find something interesting, and you chase links backwards from it. Those links, of course, are to pages that already existed when you read the first interesting thing; if those pages themselves have links, those links point even further backwards in time. It's like being trapped in a hypertext version of Merrily We Roll Along: the narrative order of your experience is the opposite of its chronological order. And this is no way to have a conversation.

Until now, the really "conversational" media on the Internet have really defied the useful arbitrary-linking structure provided by the Web. Newsgroups and email both have good notions of links to previous messages ("threading") and both have good notions of inclusion from previous messages, but both are read through interfaces that reconstruct a forward reading experience. Pretty much every solution for real conversation on the Web has involved some kind of social restriction that doesn't scale well. Discussion and community sites, on the Slashdot model, rely on a centralized server to filter comments and present them in "order." Wikis go even further: they depend on a community of editors small enough--or organized enough--to do its own ordering; they often supply a crude "inbound link" facility, but leave it up to individual editors to rationalize content at both ends of a link.

Blogs have been slamming their head against this wall for a while. I've seen two tricks in use, neither of them really satisfying. On the one hand, there is the "updated!" approach: if someone links your story with an interesting commentary, you edit your entry to not the commentary. On the other, there is the approach of linking to a blog, rather than to a specific entry. Either way, the idea is that someone following the link gets, not the entry that was there at the time of the link's creation, but instead the most recent appropriate entry. The former approach is the linkee's way of implementing this hack; the latter approach is the linker's technique.

But both these approaches run into the same problem: they undermine the "snapshot" aspect of making a hyperlink. The link no longer references an item as it existed at the time of the link. The link now references a stream that sometimes spits out items, rather than items themselves. For blogs, with their cultural esthetic of timestamped entries, this switch can be jarring. In particular, it makes it harder to follow links in their original, backward-in-time direction, since the item being linked has become a moving target. You have to quote more text when you link, because the think you're linking to might change; this increased quotation cuts directly against the idea that hyperlinks eliminate the need for extensive cut-and-pasting. Forward-in-time links, in this model, come at the cost of making all links less tight: pages and blog entries need to stand more on their own.

Trackback completely cuts this Gordian knot. There are ordinary links in blog entries: these ones take you to the thing being commented on, backwards in time. But if you Trackback-ping when you make these links, you've created Trackback links in the stories you're commenting on. These Trackback links take readers from the commented-on to the comments, forwards in time. You can run down an entire conversational tree, from start to finish, just by reading Trackback links after you read blog entries.

There's never any need to modify an entry as such. If you want to reply to someone else's comments on it, you just Trackback the entry from which they Trackbacked you. If you want to adjust your entry for everyone in response to someone else's useful suggestion, you just Trackback your own entry. You can thread Trackback trees through a single blog just as easily as through multiple blogs. (In some sense, Trackback makes the whole notions of blog categories, comments, and even sequencing of entries obsolete: all of these features could be reinterpreted as semantics layered on top of a layer of Trackback two-way links.) Just to make it all even better, there is no central server coordinating everything, so Trackback fits in nicely with the decentralized Web Way of doing things.

Trackback and the Meaning of Commentary

Trackback isn't the first technology to try to allow people browsing the web to jump from content to commentary on the content. There have been plenty of annotation schemes (usually in the form of third-party web sites or browser plug-ins) to try similar things: this space is where Alexa, among others, got its start. But Trackback is the first such technology to be both highly usable and also honestly and truly decentralized, and as such, I think it has an excellent shot at catching on, big-time.

Given this possibility, I'd like to ask how our notions of "commenting" change when we have a technology capable of linking from content to commentary across servers. For me to comment on your content--indeed, for me to say anything at all--three things have to happen. First, I have to have useful access to the right source materials (here, your content0: if I don't know about it or never find it or can't read it, I won't even be inspired to say something about it. Second, I need to have the resources to comment: I need, after all, to fix my ideas in some tangible expression, as a copyright lawyer would say. And third, I need to have marketing for my comment: people who might read your content (including, of course, you) need to know that I've commented on it and be able to get at my comment.

Let's bracket the first issue, access; it comes up both in the real world and online and if we take too zealous a line here, we'll only be creating needless privacy problems. People need to be able to say things not intended for everyone else in the world; one could write an entire essay on just one small aspect of balancing access to speech with privacy concerns. The second issue, resources, is in some sense the issue that the Internet can make vanish. After all, as storage and bandwidth have bcome cheaper, it's been easier to get online; you don't have to be rich to have a home page. The third issue is and has always been the problem: anyone can have a home page but hits are harder to come by. From spam to search engines, that contest for the limited bandwidth of human attention is the bottleneck all across the Internet.

With blogs and web pages more generally, we tend to conflate the issues of resources and marketing, because there's usually some kind of rough connection between the resources needed to support a given statement and the size of its distribution. The Slashdot Effect is just what happens when marketing suddenly outstrips resources. This conflation, though, is only going to cause us more and more intellectual difficulties as time goes on, because resources keep on getting cheaper and cheaper, while marketing is never really going to find a way to increase the supply of eyeballs in similar proportion.

I would like to argue, then, that commentary in particular--what Jack Balkin calls "glomming on"--exposes a serious rift between these two factors. Here in the blogaxy, once I have found a piece of content to comment on, I as a blogger can just throw my commentary in my blog and hit "publish" and I have no resources issue. But, of course, the only folks who'll know about my commentary are the ones who read my blog or get my RSS feed. And If I'm commenting on a story over at the Times, that audience is tiny compared with those looking at the content on which I'm commenting.

We tend to assume that this disparity is inevitable, because when we start thinking about the marketing problem involved--how do I get Times readers to learn about my commentary?--can only be solved through some kind of resource commitment at the content end. Their letters-to-the-editor page can only be so big; their message boards will need to store and serve all their reader comments. If you're going to run a Slashdot, goes this way of thinking, you need to upgrade your server with monotonous regularity.

But Trackback, I assert, explodes this way of thinking, because the commentary need never live with the content. Why should it? All the content reader (or the content server) needs to know is that there is commentary and where the commentary lives. If I'm commenting on a story at Freedom to Tinker, I just send Ed a Trackback ping, He expends the screen-space resources to note that I've commented; I expend the resources to format my commentary and to serve it up. And I have now tapped into precisely the marketing vehicle--Freedom to Tinker itself--that I should be using to advertise my commentary on Freedom to Tinker.

Trackback, Freedom of Speech, and the Right of Reply

Which is all a way of saying that the right of reply isn't as silly an idea as many people seem to think it is. The right of reply is an old concept that opponents of particular points of view need to be given the opportunity to respond to expressions of those views; it's getting a lot of recent attention because of Council of Europe attempts to extend it to electronic media. The typical assumptions behind right of reply are tied to large, real-world media: newspapers and over-the-air broadcasters. There, it's usually interpreted to mean equal time or some similar ability offer a response in the same format and location as the original.

The usual line goes that this reasoning is remarkably inappropriate online, where "location" is such a weak concept. As such distinguished folks as Declan McCullagh say, regulations imposing an online right of reply turn into serious restrictions of free speech, because they force bloggers and online journalists to express themselves in rigid formats and designs which have spaces particularly carved out for reply comments (which, after all, might be of unknown or exceptional length). It cuts deeply into the flexibility of online realms by forcing a particular reply architecture. Cass Sunstein proposed a variant of the right of reply in Republic.com and got raked over the coals especially harshly by folks making similar technical and free-speech-oriented arguments.

But in the Trackback context, such arguments have much less bite. That's because a Trackback-based "reply" can be so much less intrusive, both visually and in terms of server resources, than a traditional comment. Indeed, the right of reply makes more sense online than in traditional media, because you don't need to devote expensive air time or newsprint to the reply. You can just link, and online, a link is as good as a nod. It's the commenter who invests the resources and effort to format and serve the reply. Ed Felten lets anyone send him a Trackback ping--and bingo, he's enabled something that satisfies every reasonable moral notion of the right of reply. There's no reason why Glenn Reynolds couldn't do the same. Or Roger Ailes.

Now I don't mean for this observation to be the final word on the right of reply. There are serious scaling issues here. Enable Trackback at the Times and you'll have a massive thicket of Trackbacks. You'll have spam problems, and indecency ones, and probably intellectual property ones, and certainly a serious signal-to-noise issue. Indeed, you'll probably have a lot of trouble separating the replies protected from the "right" from the zillion comments of aggrieved bloggers. So, on the left hand, Trackback won't solve all the genuine technical objections to a universal right of reply, at least not by itself. And, on the right hand, the signal-to-noise problem isn't entirely a problem. Part of the point of the Internet is that everyone can speak. The notion, embedded in the right of reply, that certain replies are priveleged enough to be protected, looks awfully problematic in a world of blogs and Trackbacks.

But I do want to point out that the emergence of technologies like Trackback recontextualizes serious and ongoing debates about free speech and political culture. Unless we think seriously about the technological possibilities and limitations, we're going to be stuck repeating slogans that are increasingly inapplicable to the actual situation on the ground.

Trackback makes the Web more conversational. And in doing so, it responds directly to the concern that led Cass Sunstein to propose his right-of-reply scheme in the first place: his fear that the Web was becoming increasingly fragmented and politically polarized. His fear was that people wouldn't be brought into contact with opposing viewpoints and would thus gradually slide off into political echo chambers where their original prejudices merely reinforced themselves. Bloggers have been (allegedly) among the great counterexamples to this claim: they spend tons of time linking to stupid things the "other" side has said and pointing out in great detail the stupidity involved.

Trackback can take this tendency one step further: it can allow people reading a blog to find out not just whom its author thinks is stupid but also who thinks its author is stupid. This is, of course, a step requiring some courage. The first real test of Trackback as a social concept is going to be how people handle Trackback pings that they discover link back to pieces harshly criticizing them. The truly public-minded blogger approach would be to accept them, and to return the favor with another blog entry of one's own. What was that line about offensive speech and more speech, again?

It should be clear by now that when I say "Trackback" I really mean something more like "Trackback and other technologies like it that have yet to be invented." But that's another thing that's great about Trackback as a concept: it opens doors and suggests new and remarkable things yet to come. It's the kind of technology that makes you think other technologies, too. It's the kind of technology that makes you think about free speech and other wonderful things,

 
Related Links
· More about Free Expression
· News by James Grimmelmann


Most read story about Free Expression:
The Future of Virtual Kiddie Pr0n and Other Notes on Ashcroft v. Free Speech

Options

 Printer Friendly Page  Printer Friendly Page

 Send to a Friend  Send to a Friend

Threshold
  
The comments are owned by the poster. We aren't responsible for their content.

Re: Trackback and Free Speech (Score: 1)
by greglas on Wednesday, June 18 @ 09:06:23 EDT
(User Info | Send a Message) http://www.chaihana.com/pers.html
Great analysis of an interesting new development, JG. Thanks.

Fwiw tho, I'm skeptical if the right of reply idea is scaleable. When I see 10 trackback entries, I might browse them. But on the highly trafficked sites, you'd probably get 1000 urls, right?

That's the Achilles Heel, imho, in Republic.com -- just because in the Golden Age we were all getting the same newspaper didn't mean we were all reading the same newspaper. Some of us read the sports section and others read the business section. Same with trackback and the ROR. One would need to know how many people actually follow trackback links.


[ Reply to This ]


Re: Trackback and Free Speech (Score: 1)
by mako on Wednesday, June 18 @ 11:33:04 EDT
(User Info | Send a Message) http://yukidoke.org/~mako/
Trackback has the potential to be the most important thing to happen to the web in years- because it's one of the few recent developments with the genuine potential to redefine the way we think of the web. Please. To call this overstatement would be an.. er.. understatement. :) This write-up, like Lawmeme more generally, is frusteratingly blogcentric. It's important for you to remember that many, even most(?) Internet users are only marginally familiar with "blogging." Others, like me, just don't find blogs particularly novel or revolutionary. RSS feeds, which you mention, and many other form of inter-server data exchange protocols have existed for a while on the web. This same sort of thing has also played an important role in many non-web hypertext systems. Many of these systems have "bi-directional" links and full link integrity. Ted Nelson's had this decades ago. Wikis do it as you allude to. Nelson's system also had full version control so you can could actually "backtrack" through time as well. And sure, TrackBack is an open-ended transport, but there are plenty of open-ended XML transports out there. For sure, making this sort of explicit interconnections has some interesting applications and implications, which you've touched on nicely. But calling this "genius" overlooks a lot of history and applications and makes this sound more like a Movable Type commercial. Movable Type is a piece of blogging software, now available in both hosted and roll-your-own-hosting versions. It distinguishes itself from the zillion other blog tools out there mostly in the clarity of its interface and its well-articulated API for third-party extensions. It's also worth noting that it also distinguishes itself from the zillion other blog tools out there in that while free for non-commercial use, it's definitely not free/open source software and imposes a number of important limitations. Maybe I'm an old crumudgon but I'm just not nearly as impressed by this as you seem to be.


[ Reply to This ]


Re: Trackback and Free Speech (Score: 0)
by Anonymous on Wednesday, June 18 @ 13:20:30 EDT
Granted, I've never used it myself, but still I love it.

Must be wonderful, but if you can't be bothered to use it, that suggests there are some hinderances to this chaninging the nature of the WWW.


[ Reply to This ]


Too hard to use (Score: 0)
by Anonymous on Thursday, June 19 @ 15:15:16 EDT
The reason nobody uses it is that it's too hard. It requires doing a web services API call which is a totally different technology than most blog tools need to support.

What we need instead is a simple forms based interface. If I write a blog comment on your entry, I should be able to go to a form associated with your blog and enter the URL of my blog entry, then you can add a link to it in the TrackBack list.

If they'd do this they'd find many times more people using it. But MT would lose one of their competitive advantages, so they won't do it.


[ Reply to This ]


Leges humanae nascuntur, vivunt, moriuntur
Human laws are born, live, and die

Contributors retain copyright interests in all stories, comments and submissions.
Everything else copyright (c) 2002 by the Information Society Project.

This material may be distributed only subject to the terms and conditions
set forth in the Open Publication License, v1.0 or later.
The latest version is currently available at http://www.opencontent.org/openpub/.

You can syndicate our news with backend.php

Page Generation: 0.280 Seconds