People of ACM - Brad Myers

August 19, 2025

You have noted that the HCII at Carnegie Mellon was formed in 1993 with an interdisciplinary faculty. What was the impetus for founding the HCII?

Carnegie Mellon University (CMU) has always been a key center of research in the field of human-computer interaction (HCI). In fact, the name of the field originated from the book The Psychology of Human-Computer Interaction, which was authored by Stuart Card, Thomas Moran, and Allen Newell, who were all affiliated with CMU. I started at the university in 1987, and by 1993, the leadership of the School of Computer Science (SCS) recognized that research and teaching in the HCI area would require an interdisciplinary approach, combining computer science (CS), design, and social sciences such as psychology. Given the importance of the field of HCI to digital products and services and the growth of knowledge and activities in HCI such as the annual ACM SIGCHI and UIST conferences, we decided to form a new academic department that could hire and appropriately evaluate top faculty with backgrounds in CS, design, and psychology. This approach has proven quite successful, and the HCII now has about 48 faculty, with 5 who have won the ACM SIGCHI Lifetime Achievement Awards, 10 members of the ACM SIGCHI Academy, and 12 ACM Fellows. Our alumni have also been quite successful and now are leaders of HCI in industry and as university faculty around the world.

A central goal of your Natural Programming project at CMU is to understand how software developers are thinking about a problem when they are trying to develop code or design an application programming interface (API). What is the key challenge here? In your years of working in this area, are there certain insights that you’ve gleaned about how developers think that can be applied for a variety of scenarios?

After studying how programmers think and work for the past 30 years, my group has developed several insights that continue to hold true today. The most important is that HCI methods are quite relevant and helpful in studying programmers and their problems. For example, my group uses HCI methods such as contextual inquiry and analysis, interviews, questionnaires, and personas to understand developers’ real needs, along with paper and low- and high-fidelity prototypes evaluated with think-aloud protocols, heuristic evaluations, and A/B testing to understand the effectiveness of potential designs. Similarly, APIs are the interface between the developers and the functionality they are trying to use. So, good interface design principles are important for APIs as well, such as consistency, discoverability, ease of error recovery, etc.

With respect to tools to help developers, we have found that researchers’ intuitions about what might be useful may be wrong, further motivating the need for studies of programmers. We have found that code views are central to developers, and often visualizations or AI tools are most useful as navigation aids for finding the relevant code. Not surprisingly, the ability to search is key, but not just through code—search is also needed across dynamic and static call graphs, data values backwards through time, and the output of debugging tools.

Among the many recent developments within HCI, what is an especially fast-growing research area? Where is the research area heading?

It is not surprising that a hot topic now is the intersection of artificial intelligence (AI) with most aspects of HCI. At CMU, we have many faculty working on responsible AI and algorithmic bias and AI errors, as well as how to make intelligent interfaces more human-centered and controllable by people. We are also developing new methods for incorporating AI techniques into the design and evaluation of all kinds of interfaces and development processes.

You’ve done work on user interfaces for handheld devices. Smartphones, in particular, have become so much a part of our daily lives. What new HCI application related to handheld devices may become prevalent in the near future?

I do not believe that handheld devices will be replaced by AI. Whereas many of today’s exciting AI systems are based on a chat or voice interface, there are still many situations where viewing and interacting with displays is better. For example, it might be useful to ask an AI application to find hotels that match various criteria, but it will still be most useful to be able to see the results as a list and as a visualization such as on a map, allowing users to point to and filter the results. After all, the human visual system is still our fastest input mechanism.

A barrier to the envisioned future of AI working with devices’ display output and graphical user interface (GUI) is the current architecture of having independent apps, each with its own silo of data. That makes it difficult for the AI to have sufficient information about the user. For example, if I want to know which hotel is closest to a convention center where I have sufficient points for a free room, the AI would need to know information about my points in different hotel chains and the cost of hotels in points—which is currently protected information kept in the apps of the different chains. The hotel chains might not be willing to divulge this information to the AI for marketing and security reasons.

Your recent book, Pick, Click, Flick! The Story of Interaction Techniques, is a significant undertaking and has received enthusiastic reviews. What made you decide to write this book? What do you hope people will take away from it?

I have always found interaction techniques interesting. I personally have used, studied, and even invented interaction techniques since 1975. Having taught a course on interaction techniques five times over the last 11 years, it seemed useful to capture the organized knowledge about interaction techniques in a book. Although many aspects of the older, important systems have been documented, their interaction techniques often have not. For example, none of the many publications about the early Smalltalk system—which was the first to demonstrate overlapping windows—explained how its novel scroll bars worked. So, I discuss this in my book. I was also fortunate to be able to get detailed stories directly from many of the original designers while they were still alive (for example, the late Jean-Marie Hullot, Larry Tesler, and Bill Atkinson).

There are many important take-aways in my book. A key one—which I have emphasized in my courses on interaction techniques as well—is that low-level details matter. For example, when creating a new design for the scrolling interaction technique, we need to ask questions such as: How fast should the content move based on the user’s actions? How fast should it slow down? How can we make it easy for the user to scroll a few lines or hundreds of pages? These details cannot be explored in a Figma (collaborative tool for interface design) or paper prototype. They must be implemented to be tested. There are similar considerations for the design of menus, text entry, graphical editing, drag-and-drop, etc.

Another take-away is that whereas most UI designers can just use standard widgets from a library, sometimes a novel design is needed. Then, the options that have been tried in the past can provide guidance and inspiration.

Finally, I hope my book is useful for historians of technology and funders of research, since the way that interaction techniques work today (and tomorrow) is based on many years of research both at universities and companies, often supported by grants from the federal government.