I spent yesterday at the geeKyoto2008 conference in London. It was a small, intelligent affair, based around the question, We broke the world. Now what? I’m not sure that it answered the question, but perhaps that’s reasonable: if a hundred or so people could fix all the world’s woes, we might be further along the road to a solution than we currently are.

With one particular exception, most of the talks were interesting, entertaining, and thought-provoking. I particularly liked the idea of the Secular Sabbath, extracting some of the benefits from religio-cultural tradition without all that noisome mysticism; the AMEE project, which aims to quantify everything in terms of energy; and Playful Spaces. The atmosphere was overwhelmingly positive rather than dystopian, and about fun rather than hair shirts. I came away with a sense that we are at a point where we can do something to correct our species’ actions.

The conference planted some interesting ideas in my head. It helped to solidify something that I’ve been thinking about for quite a while: what are the ethical implications of our actions as programmers? What are the obvious and hidden consequences?

At a simplistic level, computers run on electricity. The more of them there are, and the more computations they perform, the more electricity they use. That doesn’t mean that computers are bad and evil of themselves, though: computers have changed the nature of work in many industries. They have facilitated improvements in efficiency, and they’ve reduced some of the need for travel. But I don’t want to address that balance here.

What has been exercising my mind is the way in which technical choices—and cultural choices that are then implemented as technical choices—can affect the development of computers and shape the way in which society interacts with them. I think that we have a responsibility to consider the consequences of those choices.

Take DRM. This introduces extra encryption and decryption to the media playing process, as well as requiring additional servers and bandwidth to manage and negotiate decryption keys. Online banking also introduces an overhead of encryption and decryption, but it provides a benefit in return: you get to keep your money. From my point of view, it’s far harder to argue that preventing people from listening to music in certain ways is good for society. If you do believe that retaining control over media is good, you still have to consider that it is a choice with consequences.

I don’t know how many kilograms of carbon dioxide are generated by DRM implementations. I don’t know what order of magnitude it is, or whether it even matters. It might be possible to measure it, though. At least then we’d know whether to worry. Actually, this was one of the emergent themes yesterday: the importance of measuring and visualising the consequences of our resource use.

But look at VCRs—or, more likely these days, PVRs. (Personal Video Recorders: those hard disk recording boxes.) These devices sit there, drawing current 24/7, so that they can spring into activity when a programme is broadcast, in order that you can then watch the programme at your leisure. They are essentially obsolete: you could, potentially, simply get the desired programme from the internet when you wanted to watch it. That’s one less always-on device in your house; multiplied across the nation, it’s a lot of energy.

Except you can’t do that. You can watch some recent programming from some channels, but it’s only available for a short time window. Go on holiday for a fortnight, and you can forget it. The decision to add download and playback restrictions to the BBC iPlayer and similar offerings has added complexity, reduced their utility and, as an unintended consequence, made owning a dedicated piece of hardware more desirable.

If we’re trying to reduce manufacturing, consumption, and waste, this not what we want to encourage. Perhaps that’s a dimension that’s missing: not just the immediate impact of our decisions, but the potential impact of those decisions on behaviour, and on second- and third-order effects.

With size comes increased responsibility. The enormous scale of Microsoft’s customer base mean that everything they do is magnified globally. Whether claims of the Vista landfill effect are true or not, each line of code is running on a huge number of machines. 88 million copies of Windows XP had been sold by late last year; I don’t know how many unlicensed copies are installed, but it must be a similar order of magnitude.

That means that, for every inefficiency in Windows, a hundred million computers could potentially be drawing more current than they need to. Again, though, I can’t quantify this. But other things matter, too: security, for example, or the function of the installed operating system over time. I’ve seen non-technical people replace their computers far sooner than they need to, simply because of a perception that the computer is running slowly or otherwise broken. Whether this is due to viruses, malware or just the oft-alleged slowdown that seems to afflict Windows after several months of use, it’s contributing to a turnover of computing hardware that’s higher than it needs to be.

That’s great if you’re involved in selling computers. It’s not so great for the planet as a whole.

I haven’t really worked through all this yet. This isn’t meant to be an anti-DRM or anti-Microsoft screed: they are just examples based on my recent experiences. In the spirit of release early, release often, and in recognition of the fact that I’ve been quiet on the blogging front of late, I thought I’d at least write something, no matter how half-formed.