Structured blogging

Scoble: "The idea is to make a movie review look different from a calendar entry."

This is a misleading way of putting it. To human readers, movie reviews might still look like book reviews, calendar entries, Flash games about fighting bananas or any other kind of Web page. What is useful is to provide internal structure and metadata that allows programs to organise the reviews more effectively. Enabling programs to break down a movie review into (say) title, cast and body content enables new, movie-centric applications, just as RSS' enabling programs to break down a Web page into stories enabled aggregation, syndication, etc.

What I'm interested in is less new presentation than new navigation. I want to post book reviews to my site and have them navigable by author and title, rather than by date. (So instead of seeing a calendar over on the right, you'd see maybe a couple of A-Z grids, which would take you to "all authors beginning with A." Or something.) I want the author's name to be a link that takes you to all my other reviews of books by the same author. I'd like to be able to make a link which takes you, via Technorati or the like, to a list of other reviews of the same book. Oh, and of course I want to make it easy for you to buy the book, unless it's The Da Vinci Code of course.

By the way, TypePad's TypeLists are a primitive implementation of this kind of feature. As well as some possibilities (such as linkage to Amazon), they show up some of the difficulties in the whole idea. Do I really have to provide a star rating for everything I review? What if the metadata authorities decide that a book review must specify genre, a concept I loathe, despise and will have no truck with? How much flexibility can we offer authors before the schema becomes so general as to be useless? (In the current draft, the simple-review schema makes everything optional except the title of the review. Consumers can't even rely on finding the name of the book or movie being reviewed. Conversely, there's no optional field for series or hero, so I can't subscribe to "all reviews of Stephanie Plum books.")

Now, none of these questions are new. The SGML folks have been battling with them for decades and have won limited victories such as Dublin Core. I was unable to find any reference to such efforts on the structuredblogging.org site, which I found rather alarming. I'm not proposing DCMI or OASIS or whoever as an arbiter of structured blogging standards -- they're too slow-moving and too centralised for the weblog world. But these are smart people confronting similar questions and trade-offs to the structured blogging issue, albeit in a different environment: you gotta think we could learn a great deal from their experiences (both positive and negative) and decisions.

Structured blogging is a tough problem, especially given the decentralised and fiercely independent nature of webloggers and communities. But it opens up a world of exciting applications based around tailored, fine-grained metadata, just as we already have a range of applications based around the coarse-grained metadata of RSS. It's a problem well worth cracking.