Loosely-coupled publishing

I've long admired the way my favorite book site, All Consuming, pulls book discussions out of thin air. For example, yesterday I posted an item pointing to my current InfoWorld column, which riffs on the themes of Michael Lewis' book, Moneyball. In that posting, I included an image of the book, and I linked the image to the Moneyball page at All Consuming. Alternatively, I could have linked the image to the Moneyball page at Amazon. Either way, I knew that All Consuming would scan my blog, find a reference to a book, and assimilate that reference into the Moneyball discussion. And sure enough, today the Moneyball page at All Consuming includes this element:

Weblogs that mentioned this book within the last week

I'm happy to report that InfoWorld.com has now dipped a toe into these waters. The template for articles now includes this element:


SMS Virus Alert
(Smart Mobs)

DVD Licensing and SCO as a Verb
(A Copyfighter's Musings)

It's made of pairs of links, the first being an InfoWorld.com story, the second being a weblog item that refers to the story. Currently the items included here are culled, by an editor, from the output of Technorati and Feedster, both of which offer views (1, 2) of InfoWorld-related blog discussion. The rule for inclusion in the culled list is, roughly, "items that do not merely cite the InfoWorld article, but say something substantive about it, and/or advance the story in some useful way".

Publication-related websites often use the TalkBalk device, whereby every article is (potentially) the root of a discussion thread. What we're trying here is more like TrackBack. Should we go all the way and let TrackBack-capable blogs ping InfoWorld.com articles directly? Perhaps. But for now, since we have more culled referral data than we are using, I'd like to see the current mechanism made context-sensitive. So, for example, this story by Maggie Biggs has been widely cited in recent days (1, 2, 3). When such references are available for an article, they ought to supersede the sitewide references on that article's page, thusly:


Blue Sky On Mars: Of course, Emacs with all of the Java tools incorporated could readily be considered an IDE.

Team Murder: Maybe it's because I rarely code anything terribly complex but I find most IDEs way too pushy.

The Robinson House: I also believe that IDEs tend to groom bad habits; clicking 'build' constantly to catch errors is the biggest time waster I've observed.

I'm sure we'll see this kind of thing evolve in coming weeks and months. For blog veterans, all this may seem obvious. But in the realm of traditional publications, the instinct has always been to try to form gated communities, not federate with independent voices. I think the loosely-coupled model is more compelling, so I'm delighted to see InfoWorld.com taking its first step in that direction.

Former URL: http://weblog.infoworld.com/udell/2003/09/24.html#a804