Continuing my quest to try to figure out a better way to consume the vast wisdoms of the internet, I have to break down the types of information online. This isn’t new turf: I’ve frequently referred to a sort of spectrum, with the X-axis roughly representing size and timeliness. Understanding this just one step towards trying to figure out a better way for me to gobble it up.
Think of the spectrum as starting on on the left with the smallest atomic unit of information on the internet, which right now is probably “The Tweet”. It’s limited by design to 144 characters. As you work to the right, the content grows in complexity, research, the time required to generate, probably the latency from specific events, and overall information contained.
(For a moment I’m going to ignore anything that merely acts to replicate something else. Which means I’m temporarily ignoring things like Reddit, Hacker News, Slashdot, and a large percentage of tweets. If you can toss that stuff out, you can simplify some stuff)
So you start at the left with an infinite stream of tweets. The vast majority of them are worthless. But soon after that you bump into email items. They are directed specifically at you. They require your explicit confirmation and possible action, and they are probably someone time sensitive. After that we find a host of web page comments, blog entries, news items and so forth. But at some point, the immediacy starts fading. It stops being “Now” and it starts being “Then”. You find a thousand word essay on something that happened 2 days ago. Words that required forethought or research. Reflections on events or information. You leave behind the headline news and knee-jerk commentary on the instant, and start working into things like Novels, Scholastic Research, or Hard Science. At the far right you’re looking at things that took years to produce. This stuff will require far more of your time to properly consume. There’s less of it overall, but if it takes 1 second to consume a tweet, it might take 1 day to consume a novel. You have to be choosier.
So you have to start trying to decide what makes sense for you to spend your time on. You need to start signal farming.The number of signals we could implement is limited only by time and imagination. Some of the signals are defined by the sources themselves: Did this tweet come from a friend? Did this email come from a co-worker? Does a recommender system determine that this Wall Street Journal Story mentions keywords that I think I might like? Was this novel written by an author that I like? Did this item come from an RSS feed that I explicitly added to my reader?
At this point we can bring back the news replicators I previously ignored. Those systems are acting as recommenders, and they are trying to help me make use of my time. They can provide us multiple signals… simply appearing on a certain page is one measure of taste. A story that appears on “Slashdot” means something. A story that appears on “MetaFilter” means something else entirely. Each system also has its own internal signals as well: number of comments posted, or up votes, or whatever. These measures aren’t really directly comparable: 100 comments on one system might be a slow story, but on another story it might be the biggest story ever. Google News and Techememe both do a good job of aggregating multiple sources covering the same item, but they do so without respect for my personal preference of sources. If you can determine that a dozen items are mostly the same item, you can weigh them together… and then select the source most appropriate for the reader.
A nugget of news that was posted on Reddit and Hacker News was also Retweeted by a guy in your Facebook friends list? There’s a damn good chance that you should read that. The signals are stacking up. In other words, these different indicators or importance nudge an item up (or perhaps DOWN!) on a Y-Axis above my X-Axis that sort of measures timeliness and scale.
Now I’m intentionally avoiding thinking about UI, but it’s easy to start asking simple questions: does it make sense for a reader to order their signals? their sources? Their actions will implicitly tell you stuff. And certain things are probably obvious: Email from Mom is higher priority than G+ ‘Share’ from a stranger. The former would likely require explicitly action, even if that was as simple as “Mark as Read” to confirm information from a critical source. But below a certain threshold, you encounter information that you actually might not need or actively not even want. People you dislike. Sites you find disreputable.
So when you start reading, the top of the item is the smallest and most urgent items… the stuff with the most signals. So you start with things like tweets from your friends and emails from your coworkers. You sort of work over the x-axis where the more time pressing bits come first, but as you make your way through the days headlines, and the pressing matters of the moment, you can start finding the larger items that will hopefully be few in number, but high in relevance… positively dripping in positive signals from other sources. So when you get there, maybe you won’t have to waste your time on a thousand word essay that had nothing to say… like some might say, this one.