Adriaan Beukers
Ole Bouman
Steward Butterfield
Ben Cerveny
Elisabeth Diller
Michael Douglas
Maya Draisin
Brian Eno
Marti Guixe
Ivo Janssen
Nathalie Jeremijenko
Winy Maas
Malcolm McCullough
Irene McWilliam
Sugata Mitra
Andre Oorebeek
Chris Pacione
Garry van Patter
Fiona Raby
Hani Rashid
Kristi van Riet
Rick Robinson
Alexander Rose
Tiffany Shlain
Bruce Sterling
Lisa Strausfeld
Jane Szita
John Thackara
Tjebbe van Tijen
Michael Waisvisz

We’re talking about the 5K Awards. Specifically about lightness and information, which is a bit trickier than you might imagine.
The 5K contest started last spring. The idea was to see what sort of websites or webpages people could come up with when all the HTML, which is the sort of general description language for putting all the pieces of the website together, all the image files, all the scripts that you use, all the style information – all that together total is less than 5K.
I guess some of you work in web design. There is a whole mess of constraints, and a lot of them are new, so they are unfamiliair and often difficult. On the one hand we have the servers we have to deal with, and then the bandwith that is available; we have the clients who are paying us to make the sites, and the users of the sites at the other end; we have HTML, of course, and then the document object model, which is a general structure of what a web page can be. We have the browsers that are available, and the various platforms that those browsers run on. We also have our ego to take care of as well, and that leaves a very small space to explore, to find the optimal solutions.
So, I am going to show a couple of examples of entries to the contest. I am not going to show any of the winners, as they have had enough exposure and they are easily available. The entries were judged on the total size, function, aesthetic appeal and the overall concept and originality. And I found out, after the contest had launched, that it actually fits into a tradition of things like the threeday novelwriting contest or the 48hour film making contest and apparantly a long, maybe 20year history of creating graphics demonstrations under 4K.
That’s enough talking about what 5K is. The first thing that it is, is 4960 bits. A bit is just an on/off state. The image on the right there is a screen of 4960 pixels, either on (white ) or off (black); and that is something I can point to and say: 5K of information. In English, as different languages are different, it is about 850 words of text, with the spaces and the punctuations; that is probably about 700 words of Dutch, and 5 or 6 words of German.
5K is also about a postagestampsized photograph. Moderatly compressed in JPEG format, which is actually a very good compression format, it throws away some of the information, makes it unavailable. It’s also – but I couldn’t make a picture of this – a third of a second in MP3 format.
Before I show a couple of the sites that were created for the contest, I will show, for the sake of comparison, a couple of major websites. This is the Nederlandse Spoorwegen – the Dutch train information site. It's 46K, which is actually pretty well optimised. Amazon.com comes in at 77K, which is about average for an ecommerce site. Then there's CNN, which is really huge. I don’t know why, this isn’t a rich multimedia experience or anything like that. This is just their regular webpage. You can see that just the Javascript that they are using is more than three times over the limit for the contest.
The first 5K example, appropriately enough, is called Size Matters. There is a small problem with this. It starts off with an incorrect statement, that two nibbles equal one bit. Actually, four bits equal one nibble, but there’s no point quibbling. This whole website is under 5K. It’s actually a pretty remarkable accomplishment. One of the spacesaving techniques is the use of the forground image from one slide as the background image for the next one. The reuse of information is very important in this. It climbs up the hierarchy in sizes: terrabyte, petabyte, exabyte and then the ones I haven’t heard of  I don’t know if they are real or not: zetabyte, and so on.
So, I've talked about 5K being about 850 words, or a picture of a certain size. There is something much more important and interesting that, which is: the number of possibilities in any message of 5K. There is actually 2 to the power of 4960 distinct possible messages in length 5K. Now, you heard me right, that’s 2 to the power of 4960, and if that doesn’t blow your f***ing mind than you don’t know enough about math. There is, by comparison, the standard model in physics that tells us there are 10 79 particles in the universe, that is about 2 to the power of 430. This is mindbogglingly huge.
Just to show you how I arrrived at that: you can think of 2 to the power of 1 is one bit off run, 2 to the power of 2: 4 distinct possibilities (you can have 00, 01, 11, 10), 2 to the power of 3 and that goes on. One thing to notice about this, is that each possibility or each note on the diagram has a number of neighbours, equal to the dimensionality of the possibility space, and each of these neighbours differs from it as little as possible – differs from its neighbours minimally. So in the corner we have 000, next to 010, and 001 and 100.
One byte looks like this. It’s called a hypercube and after this it becomes impossible to draw. So the next 40956 steps would take a while. So this is one byte and this is what we might call possibility space for one byte. You can see that there are 16 different possibilties. So the number of possibilities in 5K is fantastically huge. Now, you might imagine that if you created a look up table from each one of these possibilities to an utterance that someone might make, the lookup table would be, if you would print it out at 10 points, a couple of metres wide and than maybe a billion, billion, billion, billion miles long: an incomprehensible length; and you could actually match up one of those possibilities with anything that anyone could ever say. It’s practically infinite. One of these lines could be the whole of King James’ Bible, the whole of Shakespear. Again, by way of comparison, imagine that you have every person alive today, six billion people, each uttering x things per second, and they had been doing this for 10 billion years, which is about the length of time since the Big Bang. They would have produced that many utterances, which is just a tiny, tiny fraction of the total number of possibilities in 5K.
So, obviously something constrains the possibilities that are available for people. There isn’t that great a variety of possible entries. And if we did have something like this lookup table, we could say anything that you could possibly want to say in 5K or less. And in fact we do have something that matches up the possibilities to utterances, to images, to programming languages. Here is one example, and this is just the ASCII code for ‘hello’. There is the number in decimal, and underneath the number in binary.
There are some entries that play on this a little bit. This one is called Absolute Concept by Phillipes. I got about 1200 entries for this, and a lot of them just had one name. It’s quite a nice example, I think. And then the next one came from Korea and didn’t have a title but it had a tremendously long description. You can read it if you ever have the chance. And this dynamically translates letters to the Q byte code. They ask you to tell them about yourself and then send it in to them. And so on.
The codes that we use to represent information, are one sort of constraint that we have to the number of possibilities. Here is an example that shows a lot of the other ones. This is the code for a page that’s in my next presentation. And in it is HTML, Javascipt and style information. There are only so many ways that you can present this, it’s very precisely ordered. If you flip one character here and there it could be disastrous and destroy the whole thing.
Here is my cartoon of what the possibility space looks like. Oviously there is no way to really represent it, so this is the best I can do.
As we successively apply these different constraints, the space is reduced of course. And there are things like the letter distribution. In English that is: E is the most common letter, then T and then A and O and so on. And then there are the rules of grammar and syntax. There are the meanings of the words – what it makes sense to say, what it is good to say, what is excellent. And each of those, reduces the space that we have available. And this is what might be left over, after we apply the rules of English, the rules of syntax. It reminds me of something, that might be familiar to you: this is a footnote in the book, The Design of Everyday Things, which was written by Richard Donald, an industrial designer. And this is where I am going to leave off for this presentation, because it is one of the most fundamental insights, and one of the things that drove the creation of the 5K Contest: to see what people can do, and what role constraints play in the design of all sorts of things.
This is Stewart Butterfield's first presentation at Doors in a series of three.
Stewart's presentation #2
Stewart's presentation #3 