What is Good HCI?
So now that we’ve seen some examples of historical computer interactions, let’s talk about what it might take to make a good computer interface today. Think about the computer systems that you interact with on a daily basis. What about them makes them easy to use or difficult to use? Would your answer change if you gave that computer system to a toddler, or maybe your grandmother or an elderly person? So really good human computer interaction or good computer interfaces can depend widely based on the needs of the audience. Of course, there are some things that we can think about that make good human computer interactions. For example, we want it to be functional, it needs to do what it says it does. It needs to be reliable, it should always work the way we expect it to. It needs to have a high level of usability. So it’s something we can easily and effectively work with. It should be efficient. We don’t want to have to click 18 times to open up a single web page do we? It also should be maintainable both from the ideas of the user, it should be easy for them to keep their computer up and running. But it should also be maintainable on the software side for the developers that are building these apps. And one of the things that I think is often overlooked in the field of HCI, is it should be portable. We really want the same ideas to work across multiple different devices and even multiple different paradigms. And that’s something we’ll talk about a little bit later.
To really develop good HCI we have to follow this iterative design process where we start with an existing device, we analyze that device, we check our requirements, and then we can slowly make things better. It’s really interesting to think how computer systems have developed over time. And there’s some great YouTube videos looking at this, for example, considered the original iPhone design versus what we have in our modern iOS, or consider some of the earlier versions of Windows compared to the modern versions we use today. While those systems are very similar, we can see that there’s been a slow iterative design process that has slowly improved the design over time, based on the changing hardware and our changing understanding of good HCI.
Of course, good HCI also relates to a lot of different fields. We have to think about things like psychology, such as the fields of memory and perception. How do we remember things? How do we perceive things on our computer? We have to talk about sociology and how interacting with computer changes how we interact with other humans. We look at things like cognitive science, we look at ergonomics, how easy it is for us to use a computer system, we can think of things such as graphic design and interaction design, what colors and shapes and buttons would tell us. We can even think of things like speech language pathology, should our computer refer to itself in the first person or the third person? How should it interact with us? And we can even think of really interesting fields such as phenomenology, the study of consciousness, how do we make things on our computer such as that we are consciously aware of What our computer is trying to tell us. And we’re not just ignoring it and clicking next all the time.
So one interesting case study from the field of human computer interaction is the case study of automated adaptive instruction. Automated adaptive instruction is the idea of providing the user with instructions based on what the user has done previously, and what we think are some possible next actions of the user. We’re all familiar with our phones autocorrect system, how it can predict the next words we’re trying to enter. Gmail recently has done this as well. That’s kind of the same idea. But instead of just filling in words, it’s actually telling us how to use the computer system and how to do what we want to do next. This really came about in the 1990s as users were getting used to the personal computer, but they were becoming overwhelmed by the complexity of the technology. And companies such as Microsoft turn this into a very lucrative area of research.
And so probably the most interesting case study from adaptive instruction is the computer program known as Microsoft Bob, hopefully you’ve heard of Bob, but if not, this will be a really interesting thing to talk about. And I get really excited talking about this because when I was young, the first computer my parents bought came with Bob pre installed. And so my first computer interaction was actually with Microsoft Bob. This is what Microsoft Bob looks like. The whole idea behind Microsoft Bob was to take the desktop metaphor to the next level. And instead of having a desktop for our computer, we could mirror an entire house. And so our house would have a whole lot of different rooms, we could have the living room, where in this room we have things such as a Rolodex that represent our address book, we have a money bag that represents our finances. There’s a checkbook sitting on the table that represents our Quicken or our check our money software. And that was really the idea you would have things in your house in Microsoft Bob, that mirrors the things that you would use in the real world. And by clicking on those objects, you would opened the programs that related to that. And then of course, you could have different rooms, you could have a living room with all of this, you could have an office with different things, you could have a kid’s room with all the games in it. And so it was a really neat idea.
But of course, the automated adaptive instruction part is the little dog that we see here. This dog is called rover, and rover was there to try and help you navigate Microsoft Bob, it would offer to do things like help you find a program, go to another room, add something changed something, anything you wanted to do, you could interact with rover and rover would help you figure out how to do it. And it was this really neat idea. Unfortunately, I think Microsoft Bob really wasn’t quite ready for primetime. And graphically, it doesn’t look very good. And in fact, most computer users had gotten used to the desktop metaphor. And so the power users weren’t exactly excited about it either. And of course the biggest nail in the coffin was Bob itself wasn’t that great of a piece of software in some ways. It was kind of buggy and did crash a lot. How There are some people that really are big fans of Microsoft Bob even today. And so if you look online, there’s some great videos of people who have installed Bob even on a modern computer just to see how it works.
Of course, this did lead to some ideas that stuck around. For example, rover became the search assistant from Windows XP. And of course, this is the source of the infamous Clippy office assistant from Microsoft Office 97. Again, it was the same idea. Microsoft Office has a lot of features that most users don’t know how to do, such as mail merge. And Clippy, even though Clippy might have been a little bit annoying, was perfectly capable of helping you figure out how to do that. And so it kind of got a bad name, but it did have some neat ideas.
The big lesson learned from these assistive agents is companies were trying to teach users how to use computers, instead of trying to design computer systems that work the way users expect them to. And so in a lot of modern computer design, instead Have all of this instructive bits, you see us trying to more and more build computer systems that act like we expect them to. A great example of that is on your phone. If you scroll up at the top of a web page, it tries to refresh the webpage. This is especially notable on social media sites like Twitter or Facebook, you’re scrolling up looking for new content. And so it immediately reacts to that and does what it expects you to think it wants to do.
So of course, there are a lot of virtual assistants out there today we have Clippy. In 2011, we had the release of Siri on Apple devices. We’ve had Google and Cortana. And now we have other devices such as Alexa. And so we’re really into the smart, the age of the smart assistant where almost every home has a smart assistant in it. Almost every device has the capability of using one. And so we’re really seeing this come into its prime today. So what were the big lessons learned? Well, of course, this led to a lot of different ideas, such as searches on Google, if you’ve ever searched for something on Google, and you get that Searching? Are you sure you didn’t mean X? That’s an example of adaptive instruction. It’s telling you how to do things on Google.
We also see context clues on our touch interfaces, where if we swipe left and right, when there’s nothing there, it will still let you swipe and show you that there’s nothing over there to be able to swipe to. And so like we talked about earlier, the big legacy here is, instead of trying to change human behavior, we want to try and change the program to fit the expected behavior of the human. Or to put it more concisely put things where people expect them to be.