Linking from Other Weblogs to yours with TrackBack

by Bob DuCharme

Note that I said "linking from"—usually linking is about linking to
another web page, but TrackBack lets you link from someone else's weblog to
your web page. It's typically used to alert that weblog's readers that you've linked to it because you had something to say about it, and then someone reading that weblog can follow your trackback link to see what you had
to say. This is what makes it
an interesting development in the history of web linking: the web's lack of
two-way linking is a classic topic for Old Hypertext Guys to
complain about, but now we can do it. (Well, sort of.)



For example, I once
wrote a weblog entry titled Link Typing: Who
cares?
. Jeremy Smith posted
an entry
on his own weblog that commented on mine. To make it possible for
people reading my entry to find out about his comments, he added an entry to
the TrackBack
listing for "Link Typing: Who cares?"
that linked to his entry. By adding an
old-fashioned HTML link from his weblog entry to mine and a TrackBack link
from mine to his, he essentially created a two-way link. If I had written a
new entry responding to his comments, I could have added a TrackBack link from
the weblog entry with his comments back to my response. This can continue as a
linked "conversation," and people have written software to follow these
threads.



TrackBack began as a feature of Movable Type, an application that helps
to automate weblog creation. The process of "pinging" a weblog, or
adding a link from a weblog entry's TrackBack listing to your own
weblog or website, is simple and well-documented when you use Movable Type. The
Movable Type folks have also published the
details
of how a non-Movable Type system can implement TrackBack, but I
haven't found a good explanation of how someone not using Movable Type can use
one of these implementations. This was particularly frustrating for O'Reilly Developer webloggers, because the weblogging system here doesn't use Movable Type, but if you go to any O'Reilly Developer TrackBack page (for example, this one) and follow the "What is TrackBack?" link, it takes you to a Movable Type explanation aimed at Movable Type users.




Other weblogging tools have added the ability to ping weblogs, but you can do it without special client software. Many weblog sites,
including O'Reilly, offer a web form
interface
to a CGI script that can send a TrackBack ping linking any
weblog entry that supports TrackBack back to any other web page, but my favorite so
far is the one
at reedmaniac.com
, which documents the fields to fill out well enough that
I don't have to explain most of them here. The first field, "TrackBack Ping
URL," is the special URL that CGI scripts use to tell the weblog's server
that that entry has been pinged. For example, let's say you were going to add
a TrackBack link to the my first weblog entry. (And
please, go ahead, even if you just enter dummy text to get comfortable with TrackBack. I don't
expect to see any substantive pings for that one. If you really are going to make comments in your weblog about what I say here, ping this one instead; that's what TrackBack is for!) From that weblog entry
page, you would click "TrackBack" to go to its TrackBack
Display
page. The top of that page shows
http://alpha.oreillynet.com/cgi-bin/tb/tb.cgi/wlg_3035 as the "TrackBack URL
for this entry." That's what you would enter in the first field of the reedmaniac pinging
form
. The "Permalink URL" field at the end of the form is where you put the URL of your weblog
entry: the destination of the link created by the ping. While people usually discuss TrackBack in terms of linking one weblog entry to another, you can enter any URL you want in this last field, which means that a trackback link can link to anything, not just a weblog entry.



Some people claim that the conversations represented by TrackBack threads have replaced the comments once added to the end of weblogs. For example, if I want to make a point about something you said on your weblog, instead of adding a comment at the end, I might make my point on my own weblog and ping your entry so that you and others knew that I had done so. Despite hearing this, I've seen more comments than TrackBack pings on this weblog, and most TrackBack listings I've ever seen are empty, so this still needs to be proven to me.



I'm not really sure whether TrackBack is that popular, frankly. I thought
it was very cool how sixapart.com
posted a skeleton document with minimal content and encouraged OSCON 2002
attendees to ping it from their weblog postings about the conference, because
instead of a chunk of content with accompanying commentary, the corporately
authored table of contents
that grew around their skeleton became the primary
content. On the other hand, no one that I know of has done anything similar
since, and for OSCON 2003 Edd Dumbill and Dave Beckett used IRC instead of
TrackBack to collect contributed content into a central location. Still, TrackBack isn't much more than a year old, and as more people learn how to use it we may see more interesting uses of it.





Do you know of interesting TrackBack applications like the OSCON 2002 one or thread crawlers? Do you think TrackBack's use is growing, or has it leveled off?


2 Comments

yaksox
2003-08-09 05:33:39
on t/b
I wrote a bunch of stuff about this article and trackback at my webolog, here. I don't use trackback, or i would've done that instead.
BobDuCharme
2003-10-01 10:08:55
trackback-based publication
I just found http://www.lazyweb.org/, where the main content consists of the accumulation of trackback comments, turning the site into a table of contents of pointers to interesting unimplemented ideas. The basic idea of setting up a specialized web page where anyone can add comments has its roots in Wikis, and www.lazyweb.org was itself inspired by a comment on the LazyWeb wiki (http://iawiki.net/LazyWeb)--a comment that proved the value of the LazyWeb, because someone suggested an app, and someone else implemented it.