Communications of the ACM
Three Misconceptions About Human-Computer Interaction
I come to the field of HCI via a background in AI, having learned the hard way that good interaction design trumps smart algorithms in the quest to deploy software that has an impact on millions of users. Currently a researcher at IBM's Almaden Research Center, I lead a team that is exploring new ways of capturing and sharing knowledge about how people interact with the web. We conduct HCI research in designing and developing new interaction paradigms for end-user programming.
One thing I've noticed about my new field is that there is a lot of misconception about what exactly HCI means, an effect that I never experienced as an AI researcher. People generally understand AI. Computers that exhibit human intelligence? Maybe not in my lifetime, but the ideals and goals of the field have made their way into popular culture. You've seen Steven Spielberg's movie about a little boy robot, Isaac Asimov's three laws of robotics, and the infamous HAL 9000.
If you ask pop culture what the ideals and goals are for HCI, what do you get? It's not so clear. Even amongst computing professionals, there is still a lot of uncertainty about what exactly HCI is. I think the first step towards correcting these misperceptions is to raise awareness of how HCI is perceived by everyone else.
Myth #1: HCI equals UI design
Working for a large company such as IBM, part of my time is spent interacting with folks across the company and evangelizing the need for increased product usability. Too often, people assume that HCI is only good for positioning widgets within a window, or drawing pretty icons to replace "programmer art".
At a resume workshop last year, a senior systems researcher saw HCI on my CV and started talking about UI design on the iPhone as an example of successful HCI research.
While these surface aspects should not be ignored, HCI research covers a very broad range of topics that go all the way to the core of what a system will be used for, how people perceive its value, and how one can measure the impact the system has on actual users.
Rather than being tacked on to a product as an afterthought, HCI can make a difference throughout the lifecycle of a product.
Myth #2: HCI is a library you can include in your product
Another misconception I've heard while talking to folks in IBM's product divisions is that HCI is a set of library routines that you can include to make your product more usable/prettier/better. I have had people come to me and expect, in a 30 minute phone call, to come away with some code oralgorithms or colors that will instantly improve their product.
It's not that easy. What HCI teaches is the process of system design, not answers to specific questions. I liken it to people who come by my office, admire my gorgeous handknitted throw, and want me to make them one. I tell them if they want their own, I can teach them how to knit.
Myth #3: HCI research has a short time horizon
With all the focus on short-term improvements to existing UI, and the design and study of practical systems that already exist or will soon exist, it's easy to lose sight of the longer-term goals of HCI. However, ever since my AI days I try to make time occasionally to consider the bigger picture and ask myself why we are doing what we do. Having a long-term perspective for your work is important -- it keeps you from
slogging away at the same problem for years, without noticing that the problem has become obsolete, or technology has moved in a different direction. It keeps you relevant to the larger goals of helping humanity, which is what brought me to computer science research in the first place.
My personal vision of HCI is that we are creating new ways for people to interact with information, through smart interaction design and interfaces that unleash human creativity. Ten years from now, what will the web look like? How will people interact with it? What tools will people need to become full-fledged citizens of the new information economy?
More on this in posts to come.
Disclaimer: The views expressed here do not necessarily reflect the views
of my employer, ACM, or any other entity besides myself.
Comments
Marti Hearst
Tessa, thank you for helping clarify what is a common misconception. User-centered design is a process, not a library, and not a widget.
That said, I do think it is highly satisfying that we are now seeing technology that first appeared in CHI conferences about 10 years ago, then appearing in film (e.g., Minority Report), and now appearing in products (like the iPhone).
James Begole
These are great points. I remember the first HCI course I took (Rex Hartson) and I was also expecting to learn the secret formula for when to use a button, icon, menu and exactly how to lay them out for different circumstances. Very naive.
As long as technology changes, HCI research will be needed to understand and design the use of the technologies. But have we answered the big questions already? GUIs and the direct manipulation paradigm pretty much handle the common cases.
Jeremy Pickens
Tessa - very interesting post!
I am curious, how did you come to your understanding that good interaction design trumps smart algorithms? I believe you, but what you say runs counter to the prevailing winds of today. In the web arena, for example, the common wisdom is that smart algorithms trump interaction design. Search is designed as a black-box, smart algorithm function, with no user interactivity or insight into the process. The engine handles it all for you.
So what research or examples have led you to the opposite conclusion? I would be interested in another post from you on this topic.
(As an aside, since you work at the intersection of HCI and information, you might be interested in the upcoming HCIR [=HCI + Information Retrieval] workshop: http://thenoisychannel.com/2009/03/25/hcir-2009-a-pre-cfp/ )
Tessa Lau
> GUIs and the direct manipulation paradigm pretty much handle the common cases
On the desktop, perhaps this is true. However as computing moves off the desktop, and as we start needing to interact with larger scales of information, I believe GUIs and direct manipulation just don't scale up. Which brings us to search as an emerging paradigm for interaction, which Jeremy mentioned in his comment.
Jeremy, take a look at my recent workshop paper on "Why PBD systems fail": http://tlau.org/research/papers/pbd-fail.pdf
I agree that some smart algorithms -- search included -- are successful in part because on the intelligence of the algorithm. However I'd also argue that good interaction design is also key to success. Would Google be as successful if the results page were cluttered with extraneous information that distracted you from the search results?
I'll be writing on this more in future posts.
Displaying all 4 comments