Scannable conversation summaries

Writing effective heads, decks, and leads is one of the ways we can apply engineering principles -- modularity, information hiding -- to our written communication. Threaded discussions resist this technique, so our mailboxes and discussion boards are full of constructs like this:

I've long imagined an alternate reality where meaningful titles enrich those top-level views, enabling you to absorb the sense of a discussion at a glance, like so:

So far as I can tell, the unfortunate pattern we're stuck with is just a historical accident. Although the email protocol supports threading by unique message ID, email programs didn't always implement it correctly. So people got into the habit of echoing subject lines in order to make threads hang together. There seems to be no way to reverse the ongoing and catastrophic loss of information that results from this glitch.

Not in email as we know it, anyway. Sometimes you just have to start over. In a sense we're doing that with blogging, and one of the results I'm happy to see is that conversational summaries are becoming more useful. Look at this example from today's memeorandum:

This effect has nothing to do with memeorandum, per se. You see the same thing in any of the blog readers that build conversational clusters around common links. From an engineering perspective, those links are no different than the unique IDs in email and newsgroup messages. From a writing perspective, though, it's easy to understand that you're referring to the unique identifier of an item (i.e., its URL) rather than to its title. That makes it natural to write your own title, and all readers benefit.

It's fun to imagine next steps. Suppose your writing software made it easy to signal that you agreed or disagreed with an item you're citing, or that you're asking a question, or responding to a question, or posting a tutorial, or...

...OK, never mind, first things first. It's nice that we're finally starting to get scannable summaries.


Former URL: http://weblog.infoworld.com/udell/2005/12/20.html#a1357