HCI Theory

From Granizada

Jump to: navigation, search

I was asked to do a presentation to small groups of technical people. This is what fell out. Caveat emptor - hang on, nobody bought anything!

The goals of the Computer Science field Human Computer Interaction are summarised in the Wikipedia HCI article as follows. I've checked this summary with formal HCI literature and it seems to be a reasonably representative view:

HCI Goals


So Are the Goals Met?

I have yet to meet a computer that is receptive to my needs, or even responsive. I know that such things exist to an extent in laboratory models, and in some very specific embedded computing contexts (eg the interfaces for certain military hardware.) But in terms of a generalised interface available to people with access to ordinary computers - no. Not even close.

From this and other information I modestly conclude that, regardless of progress made, the field of HCI is in general a failure so far. There's lots of fascinating work, there's some wonderful prototypes. I accept that it might be the problems are solved but the world isn't ready to listen, but that comes down to the same thing too. I done some simple literature searchs, but there doesn't seem to be a lot of papers on Why HCI is a Failure so it probably isn't currently a vibrant subfield of HCI. Maybe that is because most of the time it simply doesn't matter?

Humans are very adaptable. I hypothesise that most bad HCI design is overcome by users who don't even notice how much better things should be, either because they have been conditioned to bad design (that could perhaps be said of most computer users) and/or they are so fast at working around the HCI problems. I do know from experience that HCI is one of the last disciplines to be brought into application development projects, if at all, and that ties up with my hypothesis. If it was very clearly a problem with economic impact then money would be found to fix it. For applications that have a small or no budget, and are not major applications, HCI is very rarely specifically funded. Looked at another way, this is in effect a reward for doing bad work - exclude HCI from funding and still succeed, so clearly HCI isn't needed.

So it seems to me that as a discipline no, HCI goals aren't met in a way that has a meaningful impact on most people. It also seems to me that I'm going to get any number of HCI specialists howling me down, which I'd love to happen if they can provide facts whilst doing so.

What is HCI About?

I've often wondered why about equal time isn't given to "Human", "Computer" and "Interaction" by HCI practicioners (and researchers too, all too often.) What I see instead is that the bulk of the time is put into approaching things from the "Computer" and "Interaction" components, with the "Human" bit coming last. Maybe that's because we've all got a body, but there's usually only one of whatever fancy piece of technology the HCI experiments are being done with! It isn't just technophilia, because HCI people often come from a non-technical background such as psychology.

I should mention again I said HCI practicitioners. People whose job it is to worry about user interfaces, for all kinds of definition of "interface".

In my view HCI should be more about humans than technology. Often users often have no incentive to even want to use a new way of doing things regardless of benefits. And the users often control budgets. Then there is the bogeyman of fashions and fads we humans are subject to, which often sees better HCI techniques abandoned for worse ones, giving happier users (because modern technology is being used) and poorer results. Not good if you care about how well HCI is doing these days :-)

In my wisdom as someone who knows nothing about HCI, I decided that current HCI goals are very bad, and HCI will find it hard to succeed while they are the main ends. So let's change them! For the length of time of a short talk anyway.

A New Unified Theory

Having decided a new set of HCI goals are needed, the other day on the way home from work I thought up an original theory of HCI. Since I don't know anything about HCI it may not be novel, good or even valid. But it is original work because I dreamt it up. My theory is a description of what HCI should be trying to achieve and how.

The nice thing about my theory is that it applies particularly well to areas of HCI I'm interested in:

  • giving computers some idea of most appropriate communication without trying to teach them to mimic human nature (which latter we know is hard)
  • user interfaces in extreme and/or high-stakes communications situations

If a computer has a sense of prioritised appropriateness it can use this knowledge to discard information available to present but clearly inappropriate. This is all about context, involving factors such as timeliness, identity of parties and other events in the knowable universe. All without frustrating a human user who truly does know better than the computer what the ultimate goal is.

The thing about extreme and high-stakes situations is that typically HCI is currently poorly done, innovation is low and commercial rewards are high. Oh yes, safety-critical too, meaning there is the potential to save, reduce harm to or badly hurt humans.

I call it Unified because that is the tag used by the most conceited theories, those that think they can pull together disparate fields under one overarching explanation.

All About the Theory

Come to the talk.

Personal tools