Instant outlining revisited

It's been quite a while since I wrote about instant outlining, a feature of Radio UserLand that's now also available in the OPML Editor based on the open source Frontier kernel. On that project's support site Dave Winer writes:

It's like Instant Messaging, but instead of just a little phrase or sentence, you're swinging whole outlines around the Internet.
Why do this? It's a different way to collaborate. I haven't worked this way myself, but others have. A recent firsthand account of collaboration on the OPML Editor project itself, posted by Dave Luebbert, does an excellent job of conveying what that experience can be.

Outlining is fundamental to the underlying Frontier substrate, but from the perspective of the OPML Editor there are two dominant modes: it's a Web-oriented outliner, and it's a blog-writing tool. Aspects of these modes can be confusing, as can the relationship between the two.

Consider this OPML file, which collects some links I captured at a conference yesterday. In the OPML Editor I can read this file from the Web, reorganize it, and save it back to the Web. (Alternatively I can edit a local copy of the raw XML in the OPML Editor or in any text editor; when I save the file, it automatically upstreams to the Web.) If you are subscribed to the file in the OPML Editor, you'll be alerted when I make changes. And either of us can link (aka include, aka transclude) that file into another OPML file to create a composite outline that pulls from different sources but behaves as a single entity.

One thing that will probably surprise people is that the included outline is read-only. If I include your outline as a subtree in mine, then expand it by double-clicking, nothing prevents me from reorganizing or annotating your stuff. But when I double-click again my changes are lost. The outliner should signal somehow that this will happen, and that foreign content must be replicated locally in order to be durably changed.

Some folks will expect a fully read-write experience, as in a Wiki. But that would make things a lot more complex, and I'd argue it wouldn't really be the right thing anyway. Instant outlining is closely related to blogging, and as such the model is one in which I control my writing space, you control yours, and we syndicate composite views.

Another surprise, for me, came when I copied my instant outline into a new editor window and published it as my OPML blog. Only the text headings appeared, none of the subordinate links showed up. OPML links and HTML links seem to involve different editing semantics. What's more, the blog-writing aspect of the tool seems to be sensitive to the number of levels in the outline, so deeper items vanish.

My expectation here was based on my experience with Marc Barrot's activeRenderer which I have used here (although evidently not lately...) to reflect an OPML outline to the Web in a way that preserves its structure and dynamic behavior. Of course since Marc's stuff is all done in Frontier's scripting language, it can surely be made to work in the OPML Editor -- and perhaps already does.

There's a long history behind all this, from Dave Winer's earliest roots in outlining, to the Frontier kernel and scripting environment, to the Manila and Radio UserLand blog publishing tools. So the OPML Editor is, necessarily, different things to different people. For bloggers who like to work with outliners, it's a way to manage and publish structured information sets -- notably, directories such as this collection of podcasting resources. For knowledge workers who think in outlines, it's a mode of collaboration that combines outline processing with syndication effects.

Both of these modes present opportunities worth pursuing -- within the context of the open source project and, potentially, in other programming and publishing environments as well.

Former URL: