Dave Winer is working on another outliner. Yea! I wasn’t a big fan of More or Thinktank (sorry Dave) not because there was anything wrong with them, but because there was a simpler, more intuitive alternative on the Mac at the time: Acta. I’ve used the outline Dave built into Frontier/Radio and it’s as close to Acta as I’ve found. I use OmniOutliner now. It’s not as sweet as Acta, but it’s pretty good. By now you’ve guessed that I have an Acta fetish. It’s true. Guilty as charged. I loved the Aunt-Sister-Daughter convention, even if it meant I couldn’t use Cmd-s to save documents any more.
Of course, I’m not suggesting that someone needs to go out and re-invent a 15-20 year old outliner. Acta was great for its time, but we live in a new time now.
I use outlines to organize my ideas, and but for some minor interface quirks, they’re all pretty good at that. I think the next step for outliners is to help me integrate other people’s ideas with mine. I’d like to see an outline combined with search. This is different from the outline linking that was possible with Frontier/Radio, kind of the world wide outline, which required mutual participation. I’d like to be able to grab other people’s ideas for my outline wherever a search engine can find them.
One of the nice features of the Frontier/Radio database was the ability cmd-click on a table name and create a new window for that table. What if I could cmd-click on an outline topic and create a search window for that topic? The text of the topic would be the initial query, although I could edit it if desired. The search window would give me a choice of search engines or databases that I could use via a plug-in architecture. The search results would be displayed in this window. I could select text from a search result and drag it into my outline as new topic, where I could annotate and edit it. The new topic would remember its source url so I could always go back to it. The search would persist. It would remember which topics I dragged to the outline and always display those as long as the sources remained available. It would also allow me to delete search results that aren’t relevant and supress them from future updates of the search.
By itself, this would be a pretty cool advance.
Cooler still would be if the search window were a little smarter. What if the search window supplemented the relevance ranking supplied by the search engine, which is generic and is calculated without information about the context? I’d like to see the search window use the text of the topic from which it was created, the text of its subtopics, the text of any search results that were dragged into the outline, and the text of any deleted search results as context information for calculating relevance. Comparing search results to the outline context from which they were sought and weighting the search engine’s relevance ranking by similarity to the outline context from which the search was launched strikes me as a very useful thing. For this to work, the search window would have to actually fetch search results in order to do the relevance calculation or the search engine would have to provide an API for doing this calculation.
Would I be greedy if I mentioned that an RSS feed of new search results from an outline would be nice too?