I’ve been meaning to write something on this over the last few days and have finally got round to sitting down with my laptop and typing. All this is stemming from the last geek dinner. It’s amazing who you can meet at these things and the ideas you can pick up.
Anyway, what’s all this attention stuff about? From reading things on the web this is my understanding of it. Part of the reason for this post is for people to point me in the right direction if I’m completely off target!
From the viewpoint I’m going to look at this stuff, attention is how much you rate something, and in this case the things are blog posts. There are many ways to work out how much attention I am giving each blog post I read. A few are:
- how long I spend reading it (taking word count as a factor)
- how many times I have read it
- whether I have followed links that are contained in the article.
- an actual user rating of 1 – 10.
Attention data can be stored in an OPML file. Opml is basically an xml schema. To be honest you can do this without OPML, but people seem to be adopting it, and if something like this is to work there needs to be standards. I’d looked at OPML before all the attention stuff and didn’t really get what the excitement was about, still don’t to a certain degree. Anyway, back to the point of this post. I now have my attention.xml file. I can share this file with people and they can read blog posts I read starting with the things I have given the most attention to. Ie they can view my attention.xml file as an newspaper, with the most important things at the front, with least important last (meaning they may discount the things at the end altogether).
To be honest people probably don’t give a dam what I give my attention to. A lot of people however may be interested in what Robert Scoble and Dave Winer are reading. At the start of the morning you could fire up your attention reader, grab Robert and Dave’s attention files, let the application merge, match, and sort the data, and then get a list of posts sorted by attention. Posts which both have given a lot of attention to will be first, posts that either have given a lot of attention to will be in the middle, with little attention posts at the end. Grab and mix in more peoples attention data and I wonder what would happen. Better results, or would you end up with personal noise near the top?
Another use I see is a kind of ‘Attention Point’, where a group of people subscribe and publish and at the same time. Bit more thought needs to go into that……
So that’s my understanding of this whole thing. There are a number of barriers to all of this though. Firstly is agreeing a spec for attention.xml. There was an attempt in 2004, but that seemed to fizzle out. December just gone, and people have started talking about it again. Perhaps this time something concrete will come out of it. However as long as you adhere to using the OPML spec the worst that would probably happen is you may have to rename a few of the attributes you are storing. Once the spec for attention.xml is sorted (if it ever is!), will there be a standard way for rss aggregators to calculate how much attention is being given to things?
Interesting blogs and posts:
Steve Gillmor – original thinker of attention (or so his blog says)
Nick Bradbury – Feed demon already has attention data in it’s opml data files, although it uses proprietary attributes. Interesting posts 1 and 2 on it.
Alex Barnet – he’s just done a bit of a dump on attention stuff, plus regularly posts on it.
Dave Winer – Opml stuff.
What/who else should I be reading?
So where does SSE come into it? Well from the brief idea I put about an ‘Attention Point’, SSE could be used to make sure when a user adds a new feed to the attention pot, all users get this update. Or when new attention data is submitted, all subscribers get it. Again SSE is new to me, so if you have any pointers or ideas I’d be keen to hear them.
Please let me know what you think of this. Am I on the right track? If you interested in this stuff as well drop me an email. Would be great to be able to discuss this stuff with others.