Blogger Con II Notes, Partie Deux


Question: What makes a blog a blog?

Answer: The form. More specifically, what distinguishes blogs from other kinds of web media, is that the granular unit of the blog, the quantum unit is the post, not the page.

This post-oriented architecture makes all kinds of things possible. It’s the backbone for the permalink culture, for one thing. For another, it lets me put any given post on this blog exists in at least three places at once: the date archive, the category archive, and the individual archive (and, in the case of recent posts, the front page).

Now all that’s pretty cool. But unfortunately, tools for manipulating posts seem pretty limited. For example, with MT, I can sort my posts by date or category, but that’s about it (one can also sort by author, but that has limited utility for a single-author blog).

Why can’t we have more and better tools for manipulating posts as objects? I have well over a thousand posts on any number of topics in this particular blog, and I imagine that other long-term bloggers have even more than I. We need better tools to deal with these unwieldy databases.

One potential application: picking out sub-narratives from the clutter of a daily weblog. A trivial example would be my running joke about trying to get into the New York Times. This could be done by providing an easy way to see what posts reference each other (maybe you could think of it like a kind of internal trackback) and then mapping them so that the relationship between those six disparate posts would be made clearer. What would be even more useful would be a tool that could cross-reference these sub-narratives: that Another Near Miss post is not only part of my international fugitive sub-narrative, but also relates to this entry from last July, and, in a somewhat more tangential way, this entry from December, and these two photographs.

Now, I know that some of you are saying “but isn’t that what categories are for?” Well, yes and no. Categories are kind of a gross tool, useful for lumping large chunks of data together. If they’re too broad (“Things I Do During The Day”) they’re fairly useless; on the other hand, if they’re too granular (“19th Century Paintings I Saw on the Second Floor of the South Wing of The Metropolitan Museum of Art”), they’re equally useless. Like Goldilocks’ porridge, they need to be just right to be useful. Moreover, sub-narratives often cross categories; the same six posts mentioned above are scattered across three different categories.


Each box above represents a blog post arranged chronologically; the colors represent different categories. This is how most readers experience blogs; a chronological series of disparate posts. There’s no obvious connection between the various posts.

But if the narratives inherent in the various posts could be made explicit, then it might look something more like this:

pretty fancy-dan graphics, eh?

The thick line indicates a narrative relationship; the thin lines are more tenuous connections.

And if we can manage our blog posts as part of narratives, then we can look at how different narratives threads intersect and interact, and, at least in theory, get a better, fuller concept of the picture that the blogger is trying to paint.

What else can we do if we have tools that treat blog posts as objects? Scott Young pointed out at dinner on Saturday night that there’s not much difference, structurally, between an email and a blog post. They both have a date, title, sender, and a discrete block of text as the main body. So perhaps some the tools that are being developed to help manage email—data mining, for example—can be brought to bear on the blog problem. Maybe someone will develop a blog tool that puts posts in tree-like hierarchies, much like how modern email applications can thread discussions.

Another possibility to explore is per-post access control. At the moment, blog tools, with the exception of LiveJournal, have no real way to control the visibility of each object (well, there is draft/publish, but that hardly counts). LiveJournal offers a binary type of access control: users there can either publish posts publicly or to their “friends” list. As far as I know, there’s no option to only publish to a particular subset of the “friends” list (a “super-friends” list?), but something like that certainly could be useful, particularly in a corporate context.

Perhaps posts could expire after a certain amount of time. This could be useful for time-limited announcements: “Due To Rain, The Company Picnic Has Been Moved To The Cafeteria”. Perhaps time-control and access-control could be combined: for a certain period of time, certain posts are only accessible by certain users; when that period of time has expired, those posts are accessible by all users.

There is no question in anyone’s mind that blogging software has been incredibly innovative and creative over the past few years. The idea that the web could be broken up into smaller chunks than just a page was really revolutionary. I think that now that this post-as-building-block model has been firmly established, the next step is going to be figuring out new ways to manage and play with these building blocks.


You can, actually publish to distinct groups of your friends on livejournal.

This is reflected on your friends page by I personally use it not to restrict friends-only posts to certain friends, but to keep my comics separate from regular livejournal entries for easy viewing.

Paul Ford has a pretty briliant database and interface for archiving and cross-referencing posts by all sorts of categories. Check it out at

You probably know of him, but I thought I would mention it.

I'm reading this way after the fact, but great article.

Leave a comment