A few years ago, before terrorism became an important and everyday topic, I was visiting a good friend of mine at Cambridge, who was doing his PhD there. He got a scholarship from a fund with Prince of Wales as a patron or something of the sort. I can’t remember the details anymore. Anyhow, he (and lots of other postgraduate students) got an invitation to a dinner with Prince Charles.

I was staying at his place and he didn’t feel like going, so he came up with a marvelous idea. I should go pretending to be him. It’s not like he knew anyone, who’d be there. Since I didn’t have a suit with me, I borrowed a wool jumper, aiming for a sort of modest, academic look, which is the sort of idiotic notion you get by watching too many movies. It was also the only preparation we did.

I ended in a room (5 x 5 meters) filled with people in suits, whose idea of mingling is talking about their thesis and mentors. Two pieces of easily obtainable information that I didn’t have. So I’ve spent the most adrenaline filled 4 hours of my life waiting for dinner and trying to hide my wool jumper covered ass from police and anyone who could uncover me (a lot of people as I’ve found out). I had no idea it would be so difficult to avoid talking to anybody.

It’s amazing what you can get away with if no one expects you to be that daft. At the end I didn’t have to find out if police would believe my story and I discovered that, if I may be immodest for a moment, a sum of two highly intelligent people can be a cretin.

So, what does this little story tell us about group intelligence?

Absolutely nothing. I was just dying for an opportunity to share it.

On the other hand, folksonomies, wikis and most of what is these days disgustingly known as Web 2.0 do. Even though I think open formats and uninhibited access to own data should be a basic right and largely form Marela’s raison d’être, I believe in the long run it’s social part of “new” web that will prove to be really revolutionary.

Nothing exemplifies this more than Wikipedia and its sister sites. It’s exploding popularity is shocking on many levels and although most articles seem to focus on its growth and statistical data, it is relation to facts and truth that most interest me.

Critics of Wikipedia often point out sample articles of dubious quality; premise being you can’t actually know which articles to trust and which not. But it’s really the supporters , who I think are even more misguided.

Critics are not wrong. It simply doesn’t matter they are right, because we don’t care about veracity all that much. Authority is very much dead.

Web hasn’t made us critical thinkers. We don’t actually question information we find, unless we disagree with it. If anything, then first ten years of web made us more sloppy, more willing to believe unchecked statements.

Paradoxically I think it’s this naïvety that was and probably still is needed for Wikipedia’s development. It simply wouldn’t grow to its current size without satisfied readers, who are in effect a form of payment for voluntary work. Without users I doubt it would ever gather critical mass to evolve perception that you can contribute even if your knowledge might be lacking.

For all its faults, Wikipedia still is a huge body of knowledge that greatly exceeds what was available to most of us not so long ago. It’s an amazing achievement and web is markedly different place now.

After a good year of using different services I’m amazed how well simple things, which at first seemed like interesting toys, work to capture group intelligence. Tags proved to be simple, flexible way for organizing and finding things and I believe tag clouds will in future be used even more for capturing zeitgeist. I’m sure we’ll get more tools and learn to use better those we already have.

They are not perfect but as with Wikipedia, it doesn’t really matter, because they don’t have to be. They don’t have to give perfect answers as long as they avoid giving really bad ones. Rare are occasions in life where we’d prefer a perfect, but long coming answer over a good enough, yet a quick one.

Of course there are exceptions, where group intelligence fails and we need an authoritative answer. New tools won’t replace old ones and neither will groups be only source of wisdom. But they won’t go away and I’d advise against ignoring them.

George Dyson spoke about how artificial intelligence would look like and would we recognize it. Maybe it’s already here and it’s us.

List of best posts used to be here, but is not anymore. If Google can't find a page, then it probably is not worth reading.

Wednesday’s conversation with Fry provoked me to think more about where and when to use AJAX. I thought I’ve got it figured it out, but I’ve discovered that my opinions don’t really match my behavior and I find this a problem.

AJAX rage of 2005 hasn’t brought many articles exploring benefits and downsides to AJAX. There were few, even very good ones, but I’d like to see more of them and less articles describing latest gimmick or service which would probably be overlooked if it didn’t use an AJAX trick or two.

I had a dentist appointment today, something I’m sure everyone looks forward to, which at least provided me with ample time to think about this. My basic premise was that we should use technology that already works and use AJAX to build on top of that.

Since basic web model has been around for a long time, is well understood and mostly works, I tried to look at places where it doesn’t or at least not as it should and came to a short list of possible criteria for choosing when to apply AJAX:

• Browsers ought to support this, but mostly can’t or don’t. Example: autosave support for textarea, which is the most requested feature among my friends.
• Browsers support it, but badly. Example: progress bar for uploads. I don’t know why you can get a usable dialog for downloads, but you’re out of luck if you want to upload.
• Connection problems (narrow bandwidth or high latency). If this is a known, but otherwise unsolvable problem, AJAX can be used to reduce ammount and frequency of traffic.

Have I forgot anything else?

I don’t think these are only valid scenarios for AJAX use, but we did discover that fast serving of well designed pages greatly reduces the need for AJAX.