Sunday, October 24, 2004

Look Deeper

Apple Design Award CubeThe guys at mekentosj really know how to look beneath the surface of stuff. Not content with winning an Apple Design Award, they set about discovering how it worked (the apparently solid metal cube glows from within when touched) by CT scanning it and reconstructing the data as 3D visualisations. Aside from everything else, it's a beautiful example of what can be done with Open Source medical visualisation software on the Mac, and indicative of the rich possibilities for serious visualisation work on the platform.

First Batch of Xserves Installed

On Friday we got the final few Xserves installed, though a few more nodes will be added when the UPS and switches are installed. They went in like a dream, problems with the rack aside, and by the 4th or 5th we'd got them going in in under 10 minutes a piece. I think I'd hire someone if I were installing 1100 though.

Thursday, October 21, 2004

Building the cluster

Looking good
Originally uploaded by sharl.
Well, we got the call this morning to start construction of the prototype cluster that's at the heart of the Creative Supercomputing project, so that's just what we did. Those Xserves go in pretty smoothly. When all 10 are finally clustered together they'll form the most significant computer in the Institution, and we hope a powerful catalyst for creative arts research. More later, of course.

Sunday, October 17, 2004

Deep Visualisation

One of the key things to emerge from our work over the past few days is the critical role visualisation will play in the Creative Supercomputing initiative. While we realised some time ago that rapid visualisation of data (motion capture data, for instance) was a means of embedding clustered computing in the world of artistic practice, it's now becoming clear that we need to go far beyond this view of visualisation as a means to a creative end. This is bound up with the need to establish a new perspective on the role of coding in arts research, and it requires a considerable mindshift to reach the point where our work is as much technology development as application. It also means exchanging competence and comfort in our current fields for a daunting learning curve and a slow process of establishing credibility in a world rightly dominated by software engineers and mathematicians. This is a difficult transition, but it makes possible the sort of serious engagement between technological advancement and artistic practice that will be necessary to make sci-art research sustainable over the long term. In practice and in the short term it means de-emphasising surface effects and examining the nature of data sets at a deep level, trading our renderers for data mining tools. I'll be exploring the rationale and impact of this in some depth in the next few days.

Thursday, October 14, 2004


We've been working today on the Creative Supercomputing project, and it's both illuminating and disturbing to see how component projects tend to distort through the the lens of the visual arts discipline. The cross-disciplinary collaboration which gave rise to some of the projects too often seems to give way to a view of the world in which everything is treated as a design problem. I'm as guilty as anyone of seeing the penetrating gaze and shaping touch of design where there may well be none, but it's clear that not everything is a visual problem. Visual comunications and software development might both be enabled by people we refer to as designers, but one is an engineering discipline, more akin to architecture than to art. The net result of this distortion is that those of us in the visual arts tend to dramatically underestimate both the importance of underlying technologies and the burden of the necessary commitment to their continual development. We view technology as an enabling tool rather than a contextualising environment; something you deploy rather than something you populate and grow. It's a serious error, and uncorrrected it condemns our sci-art collaborations to peripheral effect and irrelevance. How we escape this fate is a theme I hope to explore more fully in the Ultraparallelism paper Gregory & are due to deliver in the next two weeks.

The Phantom Supercomputer

Artist and Canadian Technorati Michelle Kasprzak has blogged the Vivid Hothaus Progression Seminar where we gave our Creative Supercomputing initiative its first public airing. What she didn't blog (hey it was a long day) was something really interesting that emerged from our presentation. After we'd waxed lyrical for 40 minutes on the joys of building a technology project in an arts establishment one canny member of the audience wondered whether we were really building a computer cluster for the arts, or whether we'd just made it up, and whether it mattered either way. Indeed it's becoming rapidly apparent here that the catalysing and political/social power of technology is as much in its promise as in the delivery. Its the notion of the Internet as a connective force, for example, that has transformed the way we think about the world as much as the computers, software, and telecommunications that it's built from. Another way of putting this is: If we didn't have a Creative Supercomputer, people would have to build one.

Wednesday, October 13, 2004

Photo iPod Redux

colour iPod?So the rumours of an imminent update to the iPod (colour screen, bigger HD, photo sync, TV output, hip flask, yada yada) are back again, quel surprise. Our considered opinion, based on what we know of Apple and some partially-informed speculation, is that (i) colour is eventually inevitable but battery life is more important, (ii) flicking through cd covers is natural and necessary but it'll take more than a screen update to make it work properly, and (iii) you don't screw around with a 92% market share.

The second point is worth expanding upon: The whole text tree structure of a 10,000 track iPod can happily sit in memory and be scrolled through in under 20 seconds. That's 500 titles a second. Now try that with 128 pixel-square images. Any acceptable scrolling experience has to at least match the 25fps or so we're used to in video, and even that's probably not high enough. Of course Apple could opt for just displaying the cover of the track that's playing, but we rather doubt that'll be considered an acceptable user experience. Give us direct digi-cam connection and bi-directional iPhoto synching however, and we'll trade our 40GB 3G in a heartbeat.

Community Culture?

Is Community the new rich media? Twelve months ago we were struggling to get students to understand the power of communities and the sheer common sense of leveraging social phenomena in online development; they were much more interested in engagability (spinning logos and splash screens as far as I can tell). Yesterday our new Web Publishing cohort presented their early project plans, and they were almost all smart social networking concepts. Of course there's a lot of blood, sweat and development between a good idea and a workable business plan, but it's heartening to see young designers finally waking up to the pointlessness of throwing multimedia experiences at disconnected individuals through a web browser.

Tuesday, October 12, 2004

Radio Usergland

I'll need that title again when I finally get around to creating my own Podcast. I'm pretty excited about the possibilities since the whole Podcasting phenomenon brings together a bunch of things I've been thinking about and discussing with colleagues: RSS feeds, iPods, targeted radio stations. There's still a huge potential for a company like Apple to come along here and really change how P2P and personal broadcasting is seen by the major labels (I'll expand on this at a later date; for now imagine every copy of iTunes turned into a legal streaming radio station through a digital certificate and a performing rights licence). Until that happens there's a window of opportunity for budding broadcasters (that'll be us) and coders to shake up both the world of radio, and the listening habits of the iPod generation.

Do you Flickr?

Originally uploaded by han solo.
Here's one of my friend Gregory's first Flickr pics, posted (and showing up in an rss feed no-less) within 5 minutes of my inviting him to see my picture stream.

If I loved Flickr any more I'd have to marry it. They've really thought clearly about the ways people might want to get photos in and out of their sites, and they even make blogging images easy. I'm impressed too by the lack-of-friction in signing up friends to the site.