XXV.5 September-October 2018
Page: 50
Digital Citation

Bridging HCI communities

Jonathan Grudin

back to top 

Computer science, design, cognitive and social sciences, human factors, information science, information systems, artificial intelligence, communication—each has an HCI/UX community. Other disciplines may also have groups sharing this focus. I have participated in many efforts to connect some of them, primarily through joint conference activity. We built bridges that some researchers and practitioners crossed to exchange information, but almost no bridge lasted more than a few years. Perhaps they served a purpose, but given the effort involved, we expected more durable structures.

back to top  Insights


We learn from bridge failures. To better understand why ours failed, I interviewed people who crossed a bridge a few times and then stopped. My goal was to aid future efforts. This year I participated in the CHI 2018 "Bridging HCI" workshop [1] focused on building ties between ACM SIGCHI and AIS SIGHCI. In reviewing the history, it occurred to me that conference bridges might be temporary for a reason. Short-term events and temporally constrained projects—think hackathons—could be more effective, drawing engaged members of multiple disciplines to tackle a problem that all consider important. They could unite to deliver a product: working software, a detailed policy, a shared resource.

back to top  Bridging Efforts

Without disciplinary interaction, we can't stand on the shoulders of predecessors or bootstrap one another. Each community reinvents the work of other communities. It seemed sensible to establish trading posts at conferences, but it hasn't worked well. For centuries, disciplines coexisted and formed subdisciplines on their boundaries: mathematical physics, physical chemistry, biochemistry. Shouldn't we know how to do this by now? Are we different?

In some ways, yes, we are different. Moore's law gave rise to the rapid emergence of new disciplines and subdisciplines that were the focus of past HCI bridging efforts. I viewed some from outside, such as the early interaction of SIGCHI and SIGGRAPH, but I will describe those in which I participated.


CHI and human factors. The first CHI conferences were co-organized by the new ACM Special Interest Group SIGCHI and the larger, established Human Factors Society. Many CHI founders had published in the human factors literature. This bridge was in place for the 1983 and 1985 CHI conferences, after which human factors participation largely ceased.

SIGCHI and SIGOIS. SIGCHI initially centered on the use of personal computers and individual productivity tools. The older ACM Special Interest Group on Office Information Systems, SIGOIS, focused on the previous generation of computers, cabinet-size minicomputers. A minicomputer supported several people at terminals using word processors, spreadsheets, email, and databases. The COIS conference pioneered research on collaboration support, hypertext, and other topics. Efforts to coalesce the OIS and CHI communities failed. Soon CHI picked up these topics and spun off the CSCW and HYPERTEXT conferences. As minicomputers went from being a major computer industry to extinction, COIS themes and participation shifted, and it morphed into today's GROUP conference.

MIS and CSCW in the 1980s. A management information systems (MIS) HCI community, centered in business schools, had formed in the late 1960s when mainframe computers replaced vacuum tube computers and applications were less confined to science and engineering. The organizational and workgroup orientation of MIS researchers with HCI interests seemed a natural match with CSCW when the latter appeared in the mid-1980s. The first two CSCW conferences had active MIS participation. Then the MIS researchers withdrew—another shortlived bridging effort.

CSCW and ECSCW. In the late 1980s, the North American CSCW and European ECSCW communities seemed natural partners, but CSCW was dominated by computer and telecommunication companies interested in supporting small groups, whereas the Europeans were interested in large systems to support organizations and government agencies. We stumbled over differences in priorities and methods, and separated. ECSCW formed the CSCW journal with no editors from North America and for four years, only Chris Neuwirth of CMU and I were on both conference program committees. In the late 1990s, we realigned: CSCW saw the significance of the organizational contexts of workgroups, ECSCW saw organizations acquiring commercial "groupware," and both saw a role for workplace ethnographies. Ten years ago they again diverged: CSCW engaged with social media, crowdsourcing, and games, while ECSCW remained focused on work. They overlap—for example, the healthcare domain attracts both. However, the split was difficult for someone who had worked on both continents and been part of both communities.

CHI and AI. Working as an HCI developer or researcher in AI groups and labs for years, I participated in some Intelligent User Interface (IUI) and other conference events intended to bridge HCI and AI. It never happened. IUI is an active community unto itself, not a forum for the broader CHI and AI communities.

HICSS, a forum for CHI, MIS, and software engineering. The annual Hawaii International Conference on Systems Sciences has at times been a unique community bridge for subcommunities of CHI, MIS, and software engineering. I organized HICSS mini-tracks and gave papers in other mini-tracks organized by CHI and MIS researchers. HICSS has endured, but these efforts were of limited duration.

CHI and design. CHI reached out to design in the 1990s by sponsoring the Designing Interactive Systems conference series, but when few professional designers participated, it settled on software design. In the early 2000s, Richard Anderson led three impressive efforts to bridge CHI and design. He brought together CHI, SIGGRAPH, and AIGA members for three successful Design of UX (DUX) conferences. His two-day Development Consortium prior to CHI 2005, which drew HCI researchers and practitioners from the graphics, design, human factors, management, and computer science communities, led to CHI efforts to recruit more participation from the others—without lasting impact. His co-editorship of Interactions heightened its emphasis on design, which has continued and is not reliant on conferences.

ACM communities consider highly polished conference papers the pinnacle for most work, whereas other fields prize journal publication.

CHI and information science. In the late 2000s, Schools of Information proliferated and spawned a successful conference series. The iSchools are polyglot, drawing faculty from all of the fields listed above. Seeing a strong potential connection between the CHI and iSchool communities, we increased CHI participation and shepherded iConference proceedings into the ACM Digital Library when I co-chaired iConference 2011. These bridges soon crumbled.

MIS and CHI in the 2000s. Organizational IT focused on internal enterprise systems until e-commerce introduced online customers and suppliers. The MIS field, now called information systems, faced the task that CHI confronted in the 1980s: designing software for discretionary users. In 2001, the Association of Information Systems formed AIS SIGHCI with the goal of building a bridge to the CHI community. Several from CHI participated, but after a few years this bridge, like others, failed. A new generation of AIS SIGHCI leaders has revived this effort with the CHI 2018 workshop mentioned above and discussed below [1].

back to top  Why the Bridges Collapsed

Some see a bridge to another community as an opportunity to proselytize and convert, but many see an opportunity for mutual learning. What gets in the way? Long after conference-bridge alliances came apart, I contacted people who had participated for a while to ask why they stopped.

A serious impediment is that ACM communities consider highly polished conference papers the pinnacle for most work, whereas other fields prize journal publication and use conferences for community building: presenting work in progress, getting feedback, and hearing about related work. Computer science remained journal oriented longer in Europe and still is in much of Asia. Human factors and MIS researchers responded to CHI and CSCW rejections in one of two ways: "They don't like what I do" or "To polish my work enough to get it accepted, I would be most of the way to a journal publication that counts much more in my discipline than any conference paper, much less one in a different field." Portfolio-oriented design disciplines also encountered withering rejection rates. Completing the disconnect, CHI and CSCW researchers who attended conferences in the other fields castigated the unpolished works-in-progress as shoddy.

People who see where another field overlaps with their own may not see where they differ in priorities, methods, and theories. For example, one field is more interested in initial user experiences, another in making routine use more efficient. Common ground exists, but expectations are not met if the differences are not understood. For example, human factors and information systems researchers who attend CHI and CSCW conferences criticize the small participant samples and poor statistical analyses, the "unscientific" thinking-aloud method, and the lack of theoretical grounding in CHI studies.

Linguistic and cultural differences are significant even among fields relying on English and the culture of science. Common terms can have markedly different meanings, leading to significant misunderstandings [2,3]. Culturally, the "man-machine" disciplines that arose in the '50s and early '60s and were focused on military and government systems alienated many in the Vietnam-era generation that populated CHI and avoided the use of male generics. Stylistic differences tripped up reviewers. For example, MIS emphasized theory, enumerated hypotheses, and created acronyms liberally. ACM reviewers were unable to overcome irritation with these conventions and felt the authors were dissing our conferences by not conforming to our conventions.

These barriers struck the people I spoke with as significant, and many remain in place. Today, another obstacle prevents fields from connecting via mutual conference participation.

back to top  The Maturity Divide

Engineers study bridge failures. They learn from them, yet bridges continue to collapse, often because new materials and production techniques introduce fresh uncertainties. Dynamic change can render experience obsolete. This magnifies a deal-breaking barrier to conference bridges: the role of conferences in maturing fields.

An emerging discipline confronts core issues. A more established discipline has reached a common understanding of its core issues, priorities, and methods. A new discipline seeing shared interests seeks to discuss foundational issues and contribute new ideas. An established field has little interest in that; its conferences report on new, specialized explorations. Two examples follow.

In 1983, CHI was an emerging field populated by psychologists and a few computer scientists. Human factors was an established field with an extensive literature. Human factors researchers could read the entire CHI literature in a day, but the reverse was not true. Not knowing the history, CHI researchers had difficulty interpreting human factors papers. We did not get the background at conferences.

In 2001, SIGCHI was decades old and AIS SIGHCI was the new field. I could read the SIGHCI literature in a day, but the CHI literature was extensive. SIGHCI was exploring the implications as the Web brought industry IT into its first contact with "end users." CHI had dealt with initial end-user contact in the 1980s and moved on. Conference attendance wasn't an effective bridge.

The gap is greater when two mature fields consider aligning. Each has a common understanding that its members rarely consciously think about. Discussions that could be fruitful won't naturally arise through joint participation at conferences where most presenters are graduate students. It is time to consider different approaches.

back to top  New Approaches to Collaborating

Groups can collaborate by working together—or by division of labor, by creating a standard interface that enables them to exchange materials without discussion. When researching a system component, I would like a predictable interface to the components that others are working on. I do not want to rely on their experimental versions; I want to plug my current version into an existing system and see what difference it makes. If I need a machine-learning component to improve my medical application, I would like basic information and an API, not detailed descriptions of the latest tests of new deep-learning algorithms.

When work in another discipline might be relevant to yours, which would you rather do: a) attend a course covering the other discipline; b) find, learn from, and work with someone in the other discipline; or c) attend research conferences to hear presentations by students in the other discipline who assume a shared understanding that you lack and are narrowly focused on experiments around the edges? I'm leading the witness, but if c) seems unpromising, the conference-attendance approach to community exchange could be doomed... unless the conference meets in Hawaii every January.

A field could create a one-day course that describes its fundamental concepts; surfaces its shared understandings, priorities, methods, and terminologies; and surveys topics that are currently actively engaged. This traveling bridge could be taken to other disciplines as pre-conference events for those interested, with other disciplines in turn appearing at its conference.

More energizing and quicker than going back to school could be option b), for which a model exists. In a hackathon, an organization sets aside two to five days for employees to work on self-selected projects in self-organized teams to produce something concrete. Teams attract people with different roles and skills. The creative energy and accomplishment of these sanctioned skunkworks projects can be impressive.

Two or more fields could agree on a significant large-scale challenge that engages a substantial group to converge for two or three days to carry it out, perhaps prior to a conference. The task could be to create a standard, draft and polish a major policy statement, or design and build a system to support a humanitarian cause. In the course of producing a tangible result, participants learn about each other. Would you consider signing on? It might have to rely on volunteers. Grant funding for large efforts is nice when available, but incentives change and groups often align only long enough to get the funding.

Most past efforts to construct enduring associations through conference bridges lasted only a few years. The hurdles of publication venues, terminology, differing priorities and cultures, and overcoming the maturity gaps wore us down. But participants learned something about each other and many acquired a sense of another discipline's potential, should a future need arise. Our expectations were too high for us to declare victory and get out, but our efforts arguably served a purpose.

The CHI 2018 "Bridging HCI" workshop organized by AIS SIGHCI emphasized the conference-bridge model but also discussed the benefits of forming multidisciplinary teams and the hackathon model. Other topics included ways to overcome submission rejection and other challenges described above. Position papers by attendees from SIGCHI and SIGHCI fields are at http://www.bridginghci.org/papers/.

New communication and collaboration tools can support effective sharing and working across groups. The pace of change in technology use makes it desirable, if not imperative, to balance the growth of specialized efforts with efforts to connect them.

back to top  References

1. Djamasbi, S., Galletta, D.F., Nah, F.F-H., Page, X., Robert Jr., L.P., and Wisniewski, P.J. Bridging a bridge: Bringing two HCI communities together. Extended Abstracts of the 2018 CHI Conference on Human Factors in Computing Systems. ACM, New York, 2018, Paper W23; https://doi.org/10.1145/3170427.3170612. See also http://www.bridginghci.org/papers/.

2. Grudin, J. Interface: An evolving concept. Communications of the ACM 36, 1 (1993), 102–111.

3. Grudin, J. From Tool to Partner: The Evolution of Human-Computer Interaction. Morgan and Claypool, 2017. Terminology confusions appear in Appendix 1.

back to top  Author

Jonathan Grudin is a principal researcher at Microsoft and an affiliate professor at the University of Washington. He is an ACM Fellow and a member of the ACM SIGCHI CHI Academy. His book From Tool to Partner: The Evolution of Human-Computer Interaction was published in 2017. jgrudin@microsoft.com

back to top 

©2018 ACM  1072-5520/18/09  $15.00

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2018 ACM, Inc.

Post Comment

No Comments Found