I’ve been thinking a lot about semantic search and what makes something relevant. I realized that, in addition to something “having significant and demonstrable bearing on the matter at hand” (M-W), there are at least two other factors that affect how useful an item of content is.
The first is: Quality. Something can be very pertinent to a topic, but if it’s unclear, incomplete, inaccurate, or just plain bad then it probably isn’t going to do you much good. A blurry image of someone doesn’t really let you know what they look like.
The second is Timeliness. This one is tricky – it has to do with the lifecycle of a content item. The most common offenders on the web are usually really old content (often with no date stamp, so you can only guess how old and out-of-date it is). But premature content can be just as worthless. How many times have you seen something of interest, months before you had reason to act on it, then when the appropriate time rolls around you’ve forgotten about it? Or maybe you read something at a time when you really didn’t understand the significance, then later you couldn’t remember where you saw it, or how to get back to it?
So, if semantics are a better way of expressing relevance, and perhaps social media sharing can help us navigate to high quality content, what’s going to help us determine the timeliness of content?
I’ve been thinking lately about information overload. It’s ironic, in a way, because I’m pretty convinced that what needs to happen in order for the semantic web to take hold is for more people, sites, and organizations to expose their data. There are a lot of control issues involved – trust, security, confidentiality, copyright – but there’s also a real danger of accelerating the general movement towards digital saturation.
I haven’t used Twhirl, but apparently it prompted Erick Schonfeld to speculate about Web 3.0′s noise cancelling powers over on TechCrunch. I’ve heard it said many times that one of the goals of the semantic web is to deal with complexity. There’s definitely a need for it, and that need is only going to get more urgent as the information-providing services multiply. Bring on the intelligent information-filtering services!
For the Semantic Technology Conference 2008, I’ll be evaluating several tools that can be used for building taxonomies. My assessment will be based more on usability and usefulness than for technical considerations. After I give the presentation (May 21st), I’ll start posting individual tool evaluations here on the blog.
Come back and check it out later. Or better yet, come out to the conference and see me present!
I don’t know what’s been going on with the Facebook application “FunWall” in the past couple days. I got about a dozen messages forwarded from people – some of them were the same message being forwarded multiple times by the same person. Some of them consisted mainly of a semi-pornographic scribble and a message telling people to forward it and see what happens.
About a month ago (was it really that long? tsk, tsk!) I went to London to speak at a one-day conference held by Henry Stewart Events. The event was organized by Madi Welend Solomon, who I met a couple years ago at the Semantic Technology Conference. There were some excellent people speaking, and I was really happy to be in their company.
It was really amazing to participate in a complete day of discussion about metadata and taxonomy. Each speaker took a different angle and addressed a different aspect of the issues and the work. The individual presentations complemented each other and came together to tell a whole story.
Here’s what some of the other participants had to say about it: