Curiouser and Curiouser!

The trouble with tagging

So I am looking at Vesper the new iPhone note-taking application from Brent Simmons, John Gruber, and Dave Wiskus. It's a functional and tasteful app, no more than you'd expect. If I didn't already have Evernote I'd probably use it regularly. As it is, it'll displace my casual use of the built in notepad.

One thing Vesper does is to allow you to add tags to your notes. I use tagging extensively in Evernote. In fact I've been using tagging a long time, I tagged my first blog post in - I think - June 2002. Having been told nobody would ever tag things I've seen tagging rise to (something like) the mainstream in apps like Flickr and Twitter.

But what seems painfully absent, so far, in these applications is any notion of what tags mean. When I tag a tweet #sotn13 it's fine as a means of grouping tweets so you can grab them in TweetDeck. But beyond simple collections isn't it valuable to connect the meaning of things?

When I started tagging I called them "topics" and my work was based on topic mapping and on formats like XTM & XFML, and on the idea that tagging systems should be connected so that meaning can be aggregated, refined, enhanced, and - most important of all - shared.

In our hubris Paolo and I even published our own spec for including topics in RSS feeds: Easy News Topics (ENT). A key aspect of ENT was the idea that topics did not stand alone but were part of a web of meaning that could be defined in many places and joined together.

I look at an app like Vesper and it's isolated, lonely, tags and wonder if this too is an idea that will be mainstream one day.

[Update]: Belatedly I realised that "trouble" worked a lot better than "problem".

A note to myself

One of the problems I am having with the current version of Curiouser and Curiouser! is something to do with character encodings. More specifically I am ending up with a Unicode sequence C2 A0 embedded in the content. When converted back to ASCII this becomes 00 A0 which is a non-breaking space. The end result is that my posts are formatted with huge lines flowing off the side of the page and breaking the layout.

Right now I have to fire up the Rails console and do something like:

    e = Entry.last
    c = e.content
    e.content = ""
    e.save!
    c = c.gsub( /\u00A0/, " " )
    e.content = c
    e.save!
    e.publish

Because this is a pain in the ass it's something I've been meaning to fix for ages. I've just never gotten around to it. Today I forgot how to fix the problem and had to figure it out all over again (much groaning ensued) so I am making this note and have instrumented the posting code to try and figure out where these characters are coming from.

Doubtless I will have the same problem with this post. In fact, I am counting on it and my logging :)

Was thinking this was a database issue but it seems like it might actually be related to the component I use for live Markdown editing/preview.

State of the Net

Last week I attended the State of the Net conference in Trieste. Trieste was lovely (but then I enjoy rain and a bit of gloom) and a great venue for a conference.

The highlight of the first day of talks was, for me, Dave Snowden's "How not to manage complexity" although Gigi Tagliapietra and Tullio Zorzet on "Why Bach?" ran a compelling second place.

The second day had a talk on 'The Personal Cloud' from Doc Searls, with a panel on VRM including Doc & Euan Semple. a panel on Quantified Self with, amongst others Adriana Lucas, and some interesting background on e-democracy from Pietro Speroni di Fenizio. There were also a couple of talks in Italian that I would love to have followed.

And this is why I loved this conference: I am a great believer that interesting things happen "at the edges". Complexity - as a science - formed at the fringes of economics, biology, and physics. Everyone has a different viewpoint and sometimes they do not seem to fit. The beauty of it is in trying to step back and see how they can fit.

It is, perhaps, an overused line but it's amazing how much the internet has changed our lives. Can you imagine such a conference 20 years ago?

I'm still reflecting on what I heard and will, I hope, write some more about the specifics.

The Amazon Marketplace is dangerous

I am finding the Amazon Marketplace to be a terribly enticing place. Perhaps it's because a lot of the books I am reading now are old(ish) but most of them are available for £0.01 + £2.79 p&p. For reading your way into a subject this is great. I've picked up 8 books on Complexity theory by Waldrop, Holland, Kauffman, Langton, et al for less than £20.

Of course there remains the question of where, precisely, I am going to put all these books. I fear I am about to descend into chaos!

Complex roots

Recently I got interested in Complexity again or, rather "complex adaptive systems". It's an idea I played with a bit after being introduced to Dave Snowden's work back in 2003. Blimey, is that really ten years ago?

The thrust of Snowden's work with Cynefin framework is about understanding that systems fall into different categories (of which chaotic and complex systems are the two most interesting and unfamiliar spaces) and that the approach you take to a system in any given configuration should be different.

What throws a decent sized wrench into affairs is that you often have no idea which category of system you are in when you're pressed to make a decision.

I'm almost at the end of Complexity by M. Mitchell Waldrop. It's an old book largely about the coming together of a collection of fascinating and incredibly bright guys in the 80's at the Santa Fe Institute where they did a lot of thinking and experimentation with complex adaptive systems.

Apart from reminding me what an ignoramus I am, and that I'm missing a lot of vital intellectual stimulation, it's also turned me back on to some interesting stuff, like cellular automata.

Following along I implemented a browser based version of Chris Langton's simple automata and then Conway's Game of Life. I also plan to have a look at Kauffman's 'genetic networks'. These are very simple worlds but still interesting to watch and making me think about some things I've played with in recent years.

In particular the idea of the emergence of complex behaviours from simple interactions. In my musical experiments, for example, I've tried to make more and more complex pieces - generally with little success. In a sense I can see this as my attempt to impose a higher-levels of order top down. But like systems burdened with many rules, the structure can become cumbersome and brittle. I'm now wondering about how I can use simple rules, interacting, to generate interesting musical expression.

I've already been interested in Genetic Algorithms and Genetic Programming and have Hollands book Hidden Order and Kauffman's At Home In The Universe queued up to read next.

So much of this stuff is from the 70's, 80's, and 90's so I feel I have decades worth of catching up to. What are the interesting problems in complexity today?

I hope it won't take me 30 years to find out.

A reader for a writer

I was neither particularly surprised, nor particularly bothered, by Google's announcement that they are killing Google Reader. I abandoned Reader some time last year and haven't missed it. Partly - I think - because I was never a sophisticated Reader user.

One of my reasons for moving away from Reader was the feeling that reading & writing should not be separate. I often hark back to the good old days of Radio Userland and an aggregator that was integrated with the authoring platform. Whatever Dave might have gotten wrong with Radio, this move was pure genius.

I've not had much time to dedicate to the project in the last month but I intend to revive some of the aggregator code I've written in the past and integrate it here into C&C. To blend the reading & writing process again.

Happy Reaktion

I'm not sure why it is but creating new instruments in Reaktor always seems to bring out my sense of fun & creativity. Even though I pretty much always build granular samplers of one kind or another the different takes on the process itself, and ways to drive them allow a lot of room for expression.

Quite often it's an idea from some other sphere that makes me think "What if?" and leads to an instrument design.

My current instrument was inspired by watching FlickrRadio. In fact I didn't find FlickrRadio that interesting in itself. The link between color & brightness of a whole image and chord notes being a little arbitrary for my taste - at least the demo video didn't really sell me on it as a musical approach.

But what it did do was to trigger my thinking about what else I could sample to drive music?

A quite obvious answer is "other music" and I hit on an idea from U&I Software's ArtMatic which allows audio to drive image generation. ArtMatic has 4 incoming audio channels that drive variables that can be used in your structure.

The A1-A4 values are created by applying eight parallel DFT filters which are summed together in pairs. A1 is the sum of filters centered at 42 and 84 Hz. A2 is the sum of filters centered at 168 and 336 Hz. A3 is the sum of filters centered at 672 and 1344 Hz. A4 is the sum of filters centered at 7688 and 5376 Hz.

Now in point of fact I've never had good results using these in ArtMatic, most likely because I am still such a neophyte at creating AM structures. But the audio input idea has always intrigued me.

So in my new Reaktor creation I am applying the same principle on a somewhat simplified basis. I have 3 input channels with a 4-pole bandpass filter on each. This allows you to tune the range of the incoming input that will drive each channel.

Then I'm using adjustable gate triggers on each channel that allow a channel to be triggered at one threshold and released at another. When a channel triggers it drives a granular sampler to start playing a loop. Each channel has it's own loop points, speed, and so forth.

I'm at the point now where the instrument is actually starting to work and make some interesting noises. And this is the really fun part for me. Grabbing an idea by the tail and trying to hold on.

Quite often my ideas shake me off before I've really gone very far. But, sometimes, I can hang on long enough to end up somewhere really interesting. Reaktor is a great tool for listening to what those places might sound like.

Bread head

I think I have just realised why (a) I seem to crave bread so much, and (b) I seem to prefer one bread over another.

The principal discovery that I've made is that my favourite store breads seem to have between 1.9-2.0g of sugar per slice. By contrast the bread I've just eaten has 1.2g per slice and it just didn't seem "as nice" (so, of course, I had to have two slices).

In the past when I've dieted I've given up bread because the calories-per-slice were out of kilter to the "value" of a slice of bread. Now I realise why.

So I'm going to have to start looking for a really low sugar bread or give it up again.

Where will I find hidden sugar next? If I find out there's sugar in brocolli I may weep.

An insightful hack

A few weeks ago @JimAnning, @Alanb, and I were sitting in Reading's Workhouse Coffee puzzling out interesting things we could work on together and hit upon the idea of a one-day experimental hack. Could we build something interesting & worthwhile, start to finish, in a day?

So on Friday we camped out at Jim's and, during the course of the day, built something we've called TubeInsight. It's kind of "crowd-sourced feedback on your YouTube video". Something you might use for a musical performance, product demo, or investment pitch that you wanted to hone with the help of friends, family, or the internet.

For a demo of what it's about watch this short video and move the slider on the left of the video up and down to reflect how you feel about different parts of it. Then see the results including the number of people who watched and a realtime graph of their aggregate reaction.

If you want to try it with a video of your own you just need to plug in the YouTube id of the video here. No sign up required. In fact nothing required except the id of the video.

It's a prototype so there's lots of directions we could take it, lots of things we haven't addressed, and lots of choices we might want to think through. For example we picked like and dislike as our axis, that might just as easily be agree/disagree, clear/confused, love/hate, or any number of other axes. We might even have more than one. And there's lots of ways we could present the information.

Interesting and worthwhile? Please drop us a line @teaserlabs and tell us. Actually we should have made this into a screencast and put it on TubeInsight...