bbtmn

“Foolproof, and incapable of error”

Saturday 21st January 2012

Whilst I was thinking about and writing the previous post, a couple of extra things came to mind which I couldn’t really fit into the post. So I thought I might as well do a follow-up with a couple of extra observations. I did intend to write this earlier, but partly I was busy (more on that in a following post) and mostly I just didn’t get around to it.

1) The previous post was not so much about aeroplanes, but more about interfaces in general. Be that with machinery like a plane, or a device like a phone, or even infrastructure or services. And it struck me that one of the few organisations that consistently manages to create things with great interfaces is Apple. Not so much with their computers (I’m really not a big fan of MacOS, probably because I’m more used to Windows), but their iOS devices (iPhones and iPads) are really good examples of things which simplify tasks through good interface design.

It strikes me that if the computing business ever starts to go slack (!), Apple could do a good business out of consultancy; imagine if they applied their UI design skills to things other than making iPhones and iPads. This isn’t as daft as it sounds; some ex-Apple employees recently set up a business to make a better thermostat. That’s a specific example of someone applying the Apple approach to interfaces to a different type of product, and I’m sure there are other things which would benefit from the same approach.

2) For some reason, I also started thinking about 2001: A Space Odyssey (spoilers follow. Although, it’s a 40-odd year old book/film, so I guess most people at least vaguely know the plot. If you don’t, then go read the book and watch the film. They’re classics). The first – obvious – point is that a lot of the interfaces in that film do appear to tend towards simplicity. There’s loads of little things: the video phone booth that Dr Floyd uses near the start of the film, the tablets that Bowman and Poole use on Discovery, all the spaceship status screens look like they’re intended to be simple, and of course there’s HAL9000

On the topic of HAL, it occurred that his demise is pretty relevant too. HAL was programmed to help the crew, to convey information to them about Discovery and about the status of the mission. But before the crew left Earth the parameters of the mission were changed; this was secret, and the crew were not to be told until Discovery reached Jupiter. As the central computer, HAL knew the real purpose of the mission, but was not allowed to tell the crew. He was being asked to hide information, to lie. This ran counter to HAL’s programming – he was designed to give information, not to hide it – and because of that conflict he perceived there to be a problem. Which he then set out to rectify…

The point is, HAL failed because the people who defined his tasks for the mission did so incorrectly. The computer carried out its tasks as it saw best, but those tasks were in conflict with each other. And so the failure of the mission was the result of misuse of the computer. Now obviously the details in this and in the example in the previous post are very different, but in general, it’s the same fault: the computers behaved exactly as they were asked, the error arose from the way people were trying to use them.

And, really, how clever is that? That 40 years ago, people were thinking about how we’ll be using these ultra-sophisticated computers, and were (in a very broad sense) predicting some of the problems that we’re starting to see. Just makes me realise how great a job Clarke (and Kubrick, I think) did in writing that story, and how many ideas they’ve managed to pack into it. I’ve read the book many times already, but I really need to re-watch the film.

Posted at 10:34 pm | Posted In: GeekMoviesTechnology Tagged:

No Comments yet. Why not write the first?