CS/UX: Contextual Information

We have tons of data. Now we need to turn it into information.

I’m first going to link you guys to two articles:

The New Information Age
and
The Interest Graph

As it turns out, both of these articles are saying the same thing: we have mountains of data, but that data is only useful if we can analyze it given an inferred context.

Now, context is pretty hard to infer, most of the time. What we’re given as input are ‘signals’, like if a user tells us what their friends are and what they like doing. But even when we want to use that example context to suggest, say, that that I make plans to go hiking with my friends Joe, Jen and Ted (with this handy-dandy coupon from Hikers’R’Us) – because we’re all friends with the same interests living nearby each other with similar spending limits and are all free at that time on Saturday, we can’t really use those signals at face value.

As it turns out, I’m not actually very good friends with Joe, but it would have been a faux paus not to add him on my friends list, since he asked. And Jen is an extreme hiker who wouldn’t let me sit on my ass half way up the trail like I want to do and Ted doesn’t actually like hiking that much, he just put it on his profile because it makes him sound interesting. See?

So that’s why this is a non-trivial problem. And that’s just the situation where we’re asking the users for their exact context. What we want our software to do is detect and guess and learn our contexts, so when we use it, it just works – and we don’t have to think about it.

But let’s assume that we can figure out what the context is. Hiking with my friends who say they like hiking isn’t such a terrible guess, anyway. We still need to apply this context to, well, everything … without overwriting the shared components of our users’ experiences (imagine if I could only find link X on google if somewhere else I had said I liked hiking). So even when context detection should be everywhere, it shouldn’t be visible everywhere.

So from the user perspective what we need is a sort of lazy push notification system. The thing we want social to solve is having to remember anything. I don’t want to remember that I wanted to go hiking in the future, I want my computer to remember. And I want my computer to remember who else likes hiking… and then I want my computer to make me better at enjoying life by motivating me to do all those things.

This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *