Note: Stanford students can use the Stanford Library proxy for off-campus access to the readings posted on ACM Portal.
|Jan 8||Class Welcome and Intro: Interaction||
The Computer for the 21st Century, Mark Weiser, Scientific American, September 1991, pp. 94-104
This is one of the most influential, highly-cited papers in all of human-computer interaction. Weiser's vision of ubiquitous computing articulated a vision in which computing recedes into the background rather than stands as a focal point of our attention. Researchers and practitioners have chased this vision since 1991. We read this article because it is the foundation upon which much of the Interaction component of the class sits. What do you think about Weiser's differentiation between calm computation that recedes into infrastructure, and simply spreading computers around everywhere in the environment? What are the implications of such a vision?
Intro: Social + Crowd Computing
—What is (And Isn't) HCI Research
Discussants: Mackenzie Leake, Eric Rawn
Labeling Images with a Computer Game, Luis von Ahn, Laura Dabbish, CHI 2004
This is a foundational article that laid much of the groundwork and excitement behind what was later termed "crowdsourcing." The idea of transforming a data labeling task into a game is extremely creative; the trick is whether the game can be designed well, or else it flops and you don't get any players or data! Many of the concepts that this paper developed, for example looking for independent agreement between players to gain confidence in a result, are now part of nearly every crowdsourcing system. What kinds of tasks do you think are, and are not, applicable with this model? What design decisions that they made in the ESP Game (e.g., paired agreement) are critical to this functioning, and which could you imagine changing?
I always assumed that I wasn't really that close to [her]: Reasoning about Invisible Algorithms in News Feeds, Motahhare Eslami et al, CHI 2015
This recent article is representative of a HCI trend focusing on human interaction with opaque algorithms and AI systems. The big surprise of this article was that so few people were aware that their feeds were algorithmically-mediated. What follow-up studies might help us better understand this phenomenon?
Discussants: Crystal Zheng
User-defined gestures for surface computing, Jacob Wobbrock, Meredith Morris, Andrew Wilson, CHI 2009
This oldie-but-goodie introduced a design technique to the literature called "elicitation studies". It is useful in its systematization of a technique for identifying a design language for a new domain. This technique has gone on to be used in many, many other domains besides surface computing, including with robotics and drones. What kinds of contexts will this work in, and when will it not?
The Toastboard: Ubiquitous Instrumentation and Automated Checking of Breadboarded Circuits, Daniel Drew, Julie L. Newcomb, William McGrath, Filip Maksimovic, David Mellis, Björn Hartmann, UIST 2016
A great example of a recent design tools paper. It focuses on low-level prototyping of electronics, and anyone who has ever breadboarded a circuit knows the pain of hooking up the circuit incorrectly. The Toastboard is a clever approach to instrumenting and visualizating the circuit state to help debug. This paper shares characteristics with other successful design tools research, for example identifying an informational shortcoming that the designer has, and filling in that information gap. What general techniques could you imagine drawing from this domain and applying to other domains?
Interaction: User Interface Technology
Discussants: Benjamin Ho
Tangible Bits: Towards Seamless Interfaces between People, Bits and Atoms, Hiroshi Ishii, Brygg Ullmer, CHI 1997
The Tangible Bits paper sits alongside Weiser's Ubiquitous Computing paper as one of the foundational works of modern HCI. It came out several years after Weiser's article, but in many ways, was the innovation that made Weiser's vision actionable. Tangible bits takes a strong stance that input and output need to be aligned in concrete physical artifacts, whereas Weiser allowed digital outputs such as screens and projectors. This paper has spawned many, many projects and even entire conferences. What tradeoffs does Tangible Bits make through its strict adherence to input/output colocation?
Zooids: Building Blocks for Swarm User Interfaces, Mathieu Le Goc, Lawrence H. Kim, Ali Parsaei, Jean-Daniel Fekete, Pierre Dragicevic, Sean Follmer, UIST 2016**Video
The Zooids paper is a recent paper from Sean Follmer's lab here at Stanford, representing a modern instantiation of the tangible bits vision. Tangible bits always suffered from it being difficult to reconfigure the atoms of the physical world. Zooids offer a solution through a swarm robotic platform. The paper was influential for helping close the "atoms" gap so that the tangible interfaces could react back in the physical space. Can you think of alternative approaches to create reconfigurable outputs? What else might you do with this platform?
Methods and Methodology
Discussants: Abdallah AbuHashem, Emma Alderton
Methodology matters: Doing research in the behavioral and social sciences, Joseph McGrath, Readings in Human-Computer Interaction 1995
Methodology Matters is an excellent overview of the plurality of methods we use in HCI. There is no single method behind HCI; we draw on multiple methods (sometimes even in the same paper!) to triangulate our understanding. The sectioned pie diagram is the most durable contribution from this paper; think about the different axes and the tradeoffs each one implies. Which of the methods in that diagram are most common today, and how (or why) might we increase the prevalence of the other ones?
Statistics Tutorial after class 6pm-7pm in Littlefield 103
Discussants: Gabby Delos Reyes, Amrita Venkatraman
Activity sensing in the wild: a field trial of ubifit garden, Sunny Consolvo et al, CHI 2008
This foundational work by Stanford faculty James Landay and colleagues is a demonstration of the core technology that's now inside activity trackers such as the Fitbit and Apple Watch. The work is influential because it was the first to demonstrate that activity tracking was 1) feasible using commodity hardware, and 2) could lead to meaningful behavior change. Many papers at the Ubiquitous Computing conference and elsewhere have extended these findings. What follow-up research systems does this result inspire for you?
CrossCheck: Toward passive sensing and detection of mental health changes in people with schizophrenia, Rui Wang et al, CHI 2014
This recent work demonstrates the modern cutting edge of commodity sensing hardware. As activity recognition has moved into industry deployments, research has begun to focus on more challenging tasks such as those in mental health, tracked here. Pay attention to the pipeline of sensing and machine learning used to track the mental health changes; this general architecture is common to many such papers. Were you surprised that these outcomes could be predicted through commodity hardware? What are the social, medical, and ethical implications of this?
Social: Social Computing
Discussants: Dylan Cunningham
Predicting Tie Strength With Social Media, Eric Gilbert, Karrie Karahalios, CHI 2009
"Hey, did you know that you can tell how close two people are, just by observing their social media activity?" In 2009, the level of accuracy that Gilbert and Karahalios were able to achieve was astounding. This paper also set a model that was commonly used by later work: drawing on offline social theory, and extending or testing it in online socio-technical systems to draw new lessons. What lessons do you draw about the social theories from their results? What implications might this paper have for the design of our social computing systems — that a system can detect how close you are with someone?
Shaping Pro and Anti-Social Behavior on Twitch Through Moderation and Example-Setting, Joseph Seering, Robert Kraut, Laura Dabbish, CSCW 2017
This recent paper demonstrates a more recent instantiation of the same genre of argument that Gilbert and Karahalios articulated in the foundational paper above. This paper focuses on the question of moderation, which has gained a lot of attention in the dynamics of social systems such as Reddit, Twitter, and Facebook. Like Gilbert and Karahalios, Seering's work draws on social theory to make predictions, then identifies a clever interrupted time series design to test those theories. What systems or design research interventions does this research inspire? What research does this inspire in your regarding longer-term influences on behavior?
Social - Crowdsourcing
Discussants: Yasmeen Jassim
He says, she says: conflict and coordination in Wikipedia, Aniket Kittur, Bongwon Suh, Bryan A. Pendleton, Ed H. Chi, CHI 2007
This foundational paper was one of the first empirical investigations of crowdsourcing and social computing systems, and also one of the first examples in HCI of what later came to be called computational social science. Whereas prior papers we read were focused on social interaction, this paper is more focused on collaboration. The paper was influential for being one of the first to use a big online dataset (e.g., the edit history of Wikipedia) to answer a central question in HCI. It also told a story that was startling at the time: that Wikipedians were spending a smaller and smaller proportion of their edits to directly creating content that readers would see, instead spending a higher and higher percentage of their time discussing in the "talk" pages. What follow-up research does this inspire, in terms of better understanding the dynamics behind this decline?
Flash Organizations: Crowdsourcing Complex Work by Structuring Crowds As Organizations, Melissa A. Valentine et al, CHI 2017
Crowdsourcing had hit a complexity limit: it required goals to be so simple and modular that the path to achieve them could be entirely pre-defined. As a result, crowdsourcing systems struggled to achieve open-ended and complex goals whose paths cannot be predefined, including design and engineering. This recent paper from Michael Bernstein and Melissa Valentine's groups at Stanford enabled coordination toward these open-ended and complex goals by designing crowdsourcing systems that draw inspiration from modern modes of coordination in organizations. These flash organizations convene on-demand domain experts rather than non-expert workers and coordinate those experts using computational organizational structures rather than microtask workflows. The systems enable the resulting groups, which the paper calls flash organizations, to engage in adaptive coordination and achieve goals previously out of reach for crowdsourcing, including product design, software engineering, and game production. What tradeoffs does this approach make? What implications exist for organizations, and for workers?
Design: Design Process
Discussants: Vivian Yang
Parallel Prototyping Leads to Better Design Results, More Divergence, and Increased Self-Efficacy, Steven P. Dow et al, TOCHI 2010
Design process matters, but because it's a reflection-in-action loop, it's hard to prove which design behaviors are crucial and which are just vestigial. To unpack this question, Dow and colleagues introduced a clever experimental protocol and strategy for measuring the effectiveness of the resulting designs. The most important thing to draw from this paper is the experimental mechanism used to gain insight into alternative formulations of the design process. The question of parallel vs. serial prototyping matters too, of course, in developing a praxis of design; however, it's the higher-level experimental setup here which resonates most. What other studies could you imagine running in this vein?
The Reflective Practitioner (Chapters 2-3), Donald Schön, 1983
This book provides foundational theory that we draw on in HCI design: what is core to the activity of design, and what does expertise in design really mean? Schon's answer — reflection in action — is a frame that has proven useful, by articulating that the designer iterates by taking an action, reflecting on the result of that action, and using that reflection to gain insight into the problem in order to plan the next action. Schon's argument that design cannot be planned and is instead a process of iteration and reflection is foundational reasoning in HCI as to why our field works the way it does, and why we teach the way we do. How does Schon's work cast a new light on the design process, and the papers we initially discussed back in the Intro to Design section?
Design: Design Tools
Discussants: Victor Chen, Tzu-Sheng Kuo
SILK: sketching interfaces like krazy, James Landay, CHI 1996
Yes, James Landay was once a Ph.D. student. In those halcyon days, he developed one of the first design tools in the modern HCI canon, known as SILK ("Sketching Interfaces Like Krazy" — you can ask him about the konvenient spelling next time you see him). James's argument that prototyping tools need to maintain low fidelity appearance continues to resonate. However, the main reason this research became a classic is because it launched an entire area of computational tools to better support interface design. James claims that many interface design tools still don't get the main idea here: that prototypes need to look prototype-y, and tools need to support the creation of prototype-y prototypes rather than beautiful mocks. What other research projects might be possible if you extended this insight?
Learning Visual Importance for Graphic Designs and Data Visualizations, Zoya Bylinskii, UIST 2017
Recent work in design tools has drawn on AI and data-driven models to support the reflection element of reflection-in-action. In particular, designers often have to imagine how people will react to their designs. This paper is one of the first to demonstrate that, within certain constraints, those reactions could be simulated effectively by an AI. Think about the other forms of feedback, and other kinds of designs, that could be supported by this kind of approach.
AI + HCI
Discussants: Anand Lalwani
Principles of mixed-initiative user interfaces, Eric Horvitz, CHI 1999
Mixed initiative interaction is foundational work at the AI+HCI intersection. It addresses a fundamental question: when should the system take action on behalf of the user, when should it ask, and when should it simply not act? Up to this point, the question had been a binary one — should we give full control to the user, or should we delegate our tasks to AI agents? Horvitz demonstrated that the answer is "both! (sometimes)", and introduced a technical approach to supporting that tradeoff. While the details of the utility-based model haven't been widely deployed, the notion of mixed-initiative interfaces is. Colloquially, mixed-initiative interaction is now used to refer to an interaction style where the system suggests a possible acceleration or delegation of the user's task. Have we gone past this model today, or not? How would you extend this model today?
The Elements of Fashion Style, Kristen Vaccaro et al, UIST 2016
While much of current discussion is about transparency and fairness of AI models, there remains an open question of how interactive systems will actually make use of these models. Most work had been focused on relatively prosaic applications such as essentially CRUD ("create-read-update-delete") or form-filling applications, e.g., the kinds of tasks that Alexa performs. This paper demonstrates applications to an area that HCI had not much considered previously: fashion. Breaking into this area offers opportunities that are much different than the traditional models of "can we accelerate this user task with AI?" What other opportunities do you see following this approach, in this domain or others?
Media and creativity
Discussants: Kate Whitney
Draco: bringing life to illustrations with kinetic textures, Rubaiat Habib Kazi et al, CHI 2014**Video
This paper is a beautiful illustration (literally?) of how computational tools can empower new types of creative expression. This research was eventually launched as a product by Autodesk, called Autodesk Sketchbook Motion. Could these authoring techniques be translated to other domains as well?
Design Principles for Visual Communication, Maneesh Agrawala, Wilmot Li, Floraine Berthouzoz, Communications of the ACM 2011
Maneesh Agrawala and his collaborators developed a compelling approach to create compelling visual communication. This article is a review of several of his group's projects, and an articulation of the general approach they take to their work. Think about: what differentiates their process from a standard HCI design process? What role is computation playing, compared to what it would take to create these graphics manually?
Discussants: Isaac Goldstein
Improving the performance of motor-impaired users with automatically-generated, ability-based interfaces, Krzysztof Gajos, Jacob Wobbrock, Daniel Weld, CHI 2008
Accessibility — creating computing technologies that are usable across a broader range of people and situations — is an area that overlaps HCI substantially. This paper on ability-based interfaces has been very impactful in the space, arguing against a model centered in dis-ability, which is typically a focus on what a person cannot do. Think about: does this approach imply different designs or research systems than a disability-based model?
Pervasive Computing and Autism: Assisting Caregivers of Children with Special Needs, Julie Kientz et al, IEEE Pervasive Computing 2007
This space broadly also includes researchers working on systems to support neurodiversity. This paper was one of the first efforts to demonstrate how computational tools could support such a class of individuals, in particular children with autism. There is of course value in thinking about this paper as a particular solution, but the correct view would be to consider it more broadly as breaking open a space of technology for autism support, or for support of neurodiverse individuals more broadly. What other projects might be possible in this vein?
Discussants: Hancheng Cao, Claire Rosenfeld
'Yours is better!': participant response bias in HCI, Nicola Dell et al, CHI 2012
"Oh wait, even user studies and evaluation might need to be rethought in ICT4D contexts?" Yup. This paper provides compelling evidence that the evaluation strategies we use in HCI are going to produce demand characteristics that will skew our responses and downstream understanding. What should we do in the face of this evidence?
Designing mobile interfaces for novice and low-literacy users, Indrani Medhi et al, TOCHI 2011
ICT4D focuses on interactive technology support for global populations, and in particular those who have less access to technology than your typical HCI practitioner. This paper is an example of how drastically we need to reconsider our understandings of usability, and its processes, when designing interfaces for users who may not be literate. Instead of focusing on the details of the design, use this paper to think broadly about what elements of HCI design should remain constant and which might need to be reconfigured like this.
Discussants: Leenah Alfalih
As We May Think, Vannevar Bush, The Atlantic Monthly, July 1945
We will discuss the intellectual inheritance of HCI in today's class, but Bush's article As We May Think is often pinpointed as the moment that catalyzed the ideas that gave rise to the field. Up to this point, computers were calculation machines used by teams in large rooms to calculate ballistics. What did it take Bush to transition that vision to one of cognitive amplification? What claims that he made still resonate, and what did he foresee? Were there assumptions he made that can no longer apply?
Direct manipulation interfaces, Edwin Hutchins, James Hollan, Donald Norman, Human-Computer Interaction 1985
Direct manipulation is a core theory of HCI. This paper explains what direct manipulation is, develops a cognitive grounding for it, and articulates the gulfs of evaluation and execution. Think about the ways in which this theory might have impacted other work in this course (e.g., Tangible Bits), ways in which the gulfs remain relevant today or could be reconsidered, and how we might augment a theory of direct manipulation in today's environment.
Discussants: Tara Iyer, Sahil Yakhmi
The Psychology of Human-Computer Interaction (Chapter 2), Stuart Card, Thomas Moran, Allen Newell, 1983
This is typically a tough reading, but it's an important one. HCI originated from a fusion of computer scientists and psychologists, and one of the early important theories was that of a cognitive model --- an engineering model of human thinking that could be applied to the design of a user interface to predict the person's behavior. The most famous of these models, the Model Human Processor, is presented here. A number of well known models such as the KLM (keystroke-level model) followed this. Consider: what role do cognitive models such as this have in modern HCI? Why do we not see them more? Are there opportunities to bring them back with a vengeance?
On distinguishing epistemic from pragmatic action, David Kirsh, Paul Maglio, Cognitive Science 1994
A second major thread of influence from cognition on design is an understanding of the role that tools and design play in influencing our cognition. This is a more cognitively-inspired flavor of the ideas in Schon's Reflective Practitioner, focusing on the ways in which we think through the artifacts we use. What questions does this paper raise about the role of HCI systems in our lives?
Project Fair (Round One) at end of class
Discussants: Ricky Grannis-Vu, Christine Chen
Distance Matters, Gary Olson, Judith Olson, Human-Computer Interaction 2000
This paper is a classic. CSCW ("computer-supported cooperative work") is a vibrant field of HCI focused on collaboration, and is the field that gave rise to social computing. (Social computing broadened CSCW out from office-based collaboration environments to user-contributed content online.) The Distance Matters paper is part of the CSCW canon. This paper raised an important question as to how effective collaboration software could really ever be, and why. Do you think the Distance Matters limit is fundamental? If so, why? If not, what could change the situation?
Beyond Being There, Jim Hollan, Scott Stornetta, CHI 1992
Another classic of CSCW. This one pairs nicely with Distance Matters, since it essentially argues both that remote collaboration will never be as good, and that it could be even better. I think the question of "does this satisfy the Beyond Being There criteria?" is a good gut check for any CSCW or social computing system. What implications do you see of this idea, or what alternatives might you pitch?
Discussants: Christina Pan
Past, present, and future of user interface software tools, Brad Myers, Scott Hudson, Randy Pausch, TOCHI 2000
A classic of the UIST (technical HCI) community. User interface software tools, at least in the way that the authors discuss them, are no longer a core research thrust. However, toolkits in support of user interface development have increased in popularity. The most important idea from this paper is the threshold-ceiling diagram. Most of the design tools that we have discussed, as well as the programming tools we will discuss today, make some implicit argument about setting a goal of either reducing the threshold, or raising the ceiling. How do the tools you rely on fall into this typology? Should we be doing more of one or the other than we see today?
Visualizing API Usage Examples at Scale, Elena L. Glassman et al, CHI 2018
This recent work ties programming tools together with data mining and social computing. In doing so, it demonstrates the kinds of insight that an API developer might gather from information about how people use are using a system. Unlike most user interfaces, APIs and programs convey structured information, so this kind of analysis is possible. What kind of future projects might be possible with a similar kind of analysis?
Discussants: Noah Katz
Information Visualization (Chapter 1), Stuart Card, Jock Mackinlay, Ben Schneiderman, Readings in Information Visualization: Using Vision to Think, 1999, pp. 1-34.**Video 1, Video 2, Book Website
Stuart Card was also the primary author of the Psychology of Human-Computer Interaction reading. Here, years later, Card and colleagues have transitioned into helping establish a field of information visualization. This article lays out the goals and the fundamentals of information visualization.
When (ish) is My Bus?: User-centered Visualizations of Uncertainty in Everyday, Mobile Predictive Systems, Matthew Kay et al, CHI 2016
A modern question in infovis: how do we communicate uncertainty in a visualization so that a lay audience can understand how to interpret it? This paper is credited as one of the first to make substantial headway on this question. What alternative approaches can you think of? Should these techniques be popularized, or be restricted to domain experts?
Critiques of HCI
Discussants: Aditya Vishwanath
Postcolonial Computing: A Lens on Design and Development, Lilly Irani, Janet Vertesi, Paul Dourish, Kavita Philip, Rebecca Grinter, CHI 2010
What are unintended consequences of the HCI design process? One possibility, as Irani and colleagues argue in this article, is that the designer exports their own notion of what is valuable or what problems need to be solved, a modern colonial impulse. What should we do in the face of this issue?
Yesterday's tomorrows: notes on ubiquitous computing's dominant vision, Genevieve Bell, Paul Dourish, UbiComp 2006
This is a nice bookend to the class because it questions some of the assumptions of the first reading we ever did: Weiser's Ubiquitous Computing article. For me, the critique itself of "this is what the future looked like as seen in 1991, not today" rings true. What would that vision look like if it were recontextualized today?
Project Fair (Round Two) at end of class
Project Presentation is due at 3:30pm on Wednesday, March 18th.
Project Paper is due at 5:00pm on Friday, March 20th