May 4, 2007 | Filed under: Software Development

Here's an extremely useful resource for my geek friends. Peter van Dijck has collected some of the best presentations available on scaling web sites. Bookmark it!

Digg this entry.

March 21, 2007 | Filed under: Software Development

The APML Workgroup, established to create a standard for storing 'attention data' is going gang-busters thanks to its chief evangelist, Chris Saad. The conversation is informed and constructive and already moving the format forward after just a few weeks. Chris just announced the latest group of workgroup members. Welcome to:

• Sam Sethi (Vecosys)
• Daniel Winterstein (ThinkTank Maths)
• Steve Williams (Programmer, Digg)
• David Cancel (Co-Founder, Compete.com)
• Richard Giles/Graeme Sutherland (Co-Founders, Scouta.com)
• Elias Bizannes (PricewaterhouseCoopers)

Digg this entry.

March 7, 2007 | Filed under: Software Development

Chris Saad asks Does Media 2.0 Scale? Well, it doesn't right now that's for sure. I spent the whole morning yesterday pushing through 10 open Firefox tabs that had accumulated over the weekend, Techmeme's latest news on RIAA too-farness, hundreds of RSS feeds and tens of important emails. It was just too much. I also fell off my chair when Robert Scoble announced that he added 712 Twitter friends to follow.

It's not suprising that the BBC is reporting a rise in 'technology addiction' - it becomes a kind of survival instinct just to keep up with the information. There's that horrible feeling that we are missing something that leads to compulsive info-mining.

I don't trust anyone else to sort the wheat from the chaff, and I don't want to miss anything so I keep going, trying to fit more and more in. But all information is not equal. Some info is just a data point. Something to lodge in the mind to connect with something else to describe a larger picture. Some info needs urgent action. Likethe first kind these get highly prioritised.

What always suffers is the kind of information that requires some deep thought. It is just not possible to skim read Freedom to Tinker's analysis of reputation systems without thinking about it for a little bit.

So today I understood something important about the Attention Economy work ahead. Processing the avalanche of information is changing how we think and not all for the good. We are becoming good filters, but poor philosophers. We are good at information retrieval and storage and not so good at the long-thought.  We need machines to become better at filtering media 2.0 - show us the important stuff, let us get into the background stuff if we have the time and let us trust that we aren't missing anything. We need time to think.

UPDATE: Chris Saad continues the conversation.

Digg this entry.

February 28, 2007 | Filed under: Software Development

What an interesting thing to be up to my neck in!

Digg this entry.

February 22, 2007 | Filed under: Software Development

APML is Attention Profiling Markup Language, a portable file format containing the essence of what we find interesting on the internet that can be used to filter the signal from the noise or to find new stuff that is exactly what you need.

Chris Saad has been doing a terrific job evangelising the format and I have been pulled into the vortex. It's a very important initiative and I am proud to join the effort to make it a standard.

Digg this entry.

February 16, 2007 | Filed under: Software Development

Technology begets technology. It was only a few years ago that my tech news each day consisted of reading the CNET News.com home page. Now I have about 100 feeds, Google News Alerts, etc. I recently needed to establish a discipline to prevent looking up a midday and wondering where the morning went.

Randal introduced me to Chris Saad and his marvellous work with Touchstone. Touchstone is a tool which filters the signal from the noise. Not only is it a very beautiful piece of software (which counts for more than some people recognise) but it is the most useful attempt at productizing the idea of Attention Data.

Attention Data is a bundle of data that somehow represents what you have been paying attention to and, therefore, what you want to pay attention to today.

Touchstone takes OPML digests of your RSS feeds and optionally looks at your hard drive periodically to determine your interests. It then shows the news that is important to you in the form of a news ticker and/or a popup message. 

We need Touchstone and other software like it because the infospace is just going to get noisier, not to mention be joined by audio visual content.

There appears to be two major difficulties to address:

• Defining 'Attention'
• Privacy

The 'Defining Attention' difficulty is two-fold. Firstly, how does the algorithm determine that a certain file existing on my PC means I am actually interested? I guess this will be one of those things that becomes a special secret like for products like Touchstone like Google's PageRank. Secondly, will it be so effective that my horizon of information is limited to only the things I am 'paying attention to' putting me on an info-treadmill, cursed to walk the same ground every day as I have found sometimes with Techmeme. 

Privacy is the big one though, and this is presumably why the Attention Trust is making such a big deal about it. This needs to be more than words and rhetoric though. It is one of those things that must abide by Lawrence Lessig's idea that 'Code is law'. From what I can see of the Attention Trusts Recorder, it sends up entire histories of websites you have browsed to. Now real people can look at that data and do stuff with it. I just have to trust them. I think the code needs to make it impossible for people to violate that trust. Like Touchstone is doing, the attention data needs to stay on the users machine and be owned and explicitly deployed by them alone.

I liked this concept from pc4Media:

"We can make them pay for it, lease it, scream for it "show me the money", barter for it, whatever. The important point is that we get to decide who has access to it, how long they have access to it, and what we want in return."

But even then, there is a little bundle of data that crystalises everything that is important to me live on the Internet when that data is deployed. How does the user stay in control?

Can anyone point me to some useful sources to dig into these things more deeply?

Digg this entry.