real-world ubicomp deployments: lessons learned

3
GUEST EDITORS’ INTRODUCTION 1536-1268/06/$20.00 © 2006 IEEE Published by the IEEE CS and IEEE ComSoc PERVASIVE computing 21 Real-World Ubicomp Deployments: Lessons Learned A mong the many branches of computer science, ubiquitous computing enjoys an unusually distinguished history of creating and deploying prototypes. Why is this? A tempting answer is that many ubicomp researchers and practitioners have backgrounds in subjects such as HCI and systems—areas with a strong focus on learning from deploying work- ing prototypes. However, a more compelling answer might be that ubicomp, unlike almost every other sub- specialty with the exception of HCI, is about embedding com- puting into existing human systems. Why deployment is essential You might argue that deploy- ment is not a very cost-effective way of evaluating specific sys- tem aspects. Jesper Kjeldskov and his colleagues claim that “the added value of conduct- ing usability evaluations in the field is very little” and that “recreating central aspects of the use context in a laboratory setting enables the iden- tification of the same usability problem list.” 1 But deployment lets you observe people using the system as part of their daily lives, making it an important step toward the long-term goal of developing widely adopted ubicomp technology. Only through deployment can we learn about unexpected problems that might be critical in real systems. Indeed, many of the best-known ubicomp proj- ects are characterized by their focus on deploy- ment—including Parctab, 2 Active Badge, 3 and Classroom 2000. 4 Furthermore, some of the most valuable lessons in ubicomp have come not from designing or implementing particular systems but from the experience of trying to move those systems out of the lab and into real settings. In this issue In assembling this issue, we tried to collect important lessons researchers have learned from deployments. Such experience reports face chal- lenges in being accepted to selective conferences, because they often fail to present any single “deep” contribution. Instead, they tie together lessons that touch on many broad areas. The ethos of IEEE Pervasive Computing is to act as a catalyst for future research, so we view it as the ideal forum for presenting this type of work. In that spirit, we’re pleased to bring you four articles that present lessons learned the hard way—from large-scale deployments of impor- tant research prototypes. In “Moving Out of the Laboratory: Deploy- ing Pervasive Technologies in a Hospital,” Armando Fox University of California, Berkeley Nigel Davies Lancaster University Eyal de Lara University of Toronto Mirjana Spasojevic Yahoo! William Griswold University of California, San Diego

Upload: a-fox

Post on 10-Feb-2017

218 views

Category:

Documents


4 download

TRANSCRIPT

G U E S T E D I T O R S ’ I N T R O D U C T I O N

1536-1268/06/$20.00 © 2006 IEEE ■ Published by the IEEE CS and IEEE ComSoc PERVASIVE computing 21

Real-World Ubicomp Deployments: Lessons Learned

A mong the many branches of computer science, ubiquitous computing enjoys an unusually distinguished history of creating and deploying prototypes. Why is

this? A tempting answer is that many ubicomp researchers and practitioners have backgrounds in subjects such as HCI and systems—areas with a strong focus on learning from deploying work-ing prototypes. However, a more compelling

answer might be that ubicomp, unlike almost every other sub-specialty with the exception of HCI, is about embedding com-puting into existing human systems.

Why deployment is essential

You might argue that deploy-ment is not a very cost-effective way of evaluating specifi c sys-tem aspects. Jesper Kjeldskov and his colleagues claim that “the added value of conduct-

ing usability evaluations in the fi eld is very little” and that “recreating central aspects of the use context in a laboratory setting enables the iden-tifi cation of the same usability problem list.”1 But deployment lets you observe people using the system as part of their daily lives, making it

an important step toward the long-term goal of developing widely adopted ubicomp technology. Only through deployment can we learn about unexpected problems that might be critical in real systems.

Indeed, many of the best-known ubicomp proj-ects are characterized by their focus on deploy-ment—including Parctab,2 Active Badge,3 and Classroom 2000.4 Furthermore, some of the most valuable lessons in ubicomp have come not from designing or implementing particular systems but from the experience of trying to move those systems out of the lab and into real settings.

In this issueIn assembling this issue, we tried to collect

important lessons researchers have learned from deployments. Such experience reports face chal-lenges in being accepted to selective conferences, because they often fail to present any single “deep” contribution. Instead, they tie together lessons that touch on many broad areas. The ethos of IEEE Pervasive Computing is to act as a catalyst for future research, so we view it as the ideal forum for presenting this type of work. In that spirit, we’re pleased to bring you four articles that present lessons learned the hard way—from large-scale deployments of impor-tant research prototypes.

In “Moving Out of the Laboratory: Deploy-ing Pervasive Technologies in a Hospital,”

Armando FoxUniversity of California, Berkeley

Nigel DaviesLancaster University

Eyal de LaraUniversity of Toronto

Mirjana SpasojevicYahoo!

William GriswoldUniversity of California, San Diego

22 PERVASIVE computing www.computer.org/pervasive

REAL-WORLD DEPLOYMENTS

Thomas Riisgaard Hansen, Jakob E. Bardram, and Mads Soegaard describe their work in creating ubicomp sys-tems for hospitals. Researchers have long viewed healthcare as perhaps the most challenging deployment domain for ubicomp, because patient health is at risk when procedures or systems fail, patient privacy is an increasingly sensi-tive issue, and the day-to-day demands on healthcare professionals are already so intense that introducing any new technology is potentially very disrup-tive. This article discusses a system that has successfully navigated these

obstacles and is being used in a hospi-tal. The authors clearly highlight the value of such deployments, observing that “issues that seem trivial in the lab-oratory might become major obstacles when deploying systems in real-world settings.” The authors even distill some of their lessons into a useful checklist that researchers planning a deployment can use to smooth the transition out of the lab.

In “Practical Lessons from Place Lab,” Jeffrey Hightower, Anthony LaMarca, and Ian E. Smith present a retrospec-tive on the Place Lab project. As both a

framework for location-aware applica-tions and a collection of specifi c useful techniques to plug into that framework, Place Lab has been one of the most infl u-ential ubicomp projects in its domain. Indeed, few researchers in the field haven’t studied or used the Place Lab system. In this article, the authors focus on practical lessons from deploying Place Lab to a very large user base. Whereas most published Place Lab papers focus on specifi c technical aspects of the proj-ect, this article provides a valuable com-plement by presenting a broader view of the deployment and the many kinds of lessons learned—technical, logistical, administrative, and even social—from doing so. These lessons should be of interest to anyone with aspirations to create a widely used piece of ubicomp research infrastructure.

Practical lessons are also the theme of “Public Ubiquitous Computing Systems: Lessons from the e-Campus Display Deployments,” by Oliver Storz, Adrian Friday, Nigel Davies, Joe Finney, Corina Sas, and Jennifer Sheridan. The authors describe three separate installations of an evolving situated public display sys-tem—the operative term being “pub-lic.” Whereas the systems described in the fi rst two articles were deployed to selected segments of users (healthcare professionals and ubicomp researchers respectively), the e-Campus displays are used by and visible to members of the general public. Indeed, one of the deployments was part of an exhibition that attracted over 1,700 visitors. The authors distill their experiences from these public installations as 13 lessons that you can use as a checklist for pub-lic deployments.

In the last article, “Supporting Colo-cated Interactions Using RFID and Social Network Displays,” Shin’ichi Konomi, Sozo Inoue, Takashi Kobayashi, Masashi Tsuchida, and Masaru Kitsuregawa make interesting observations about user

the AUTHORS

Armando Fox is a research scientist at the University of California, Berkeley, and a co-founder of the newly-formed Berkeley RAD Lab (Laboratory for Reli-able Adaptive Distributed Systems). His research interests include the design of robust Internet-scale software infrastructure, particularly as it relates to the support of mobile and ubiquitous computing, as well as user-interface issues related to mobile and ubiquitous computing. He received his PhD from the University of California, Berkeley. He’s a member of the ACM and IEEE Com-puter Society. Contact him at [email protected].

Nigel Davies is a professor of computer science at Lancaster University and an adjunct associate professor of computer science at the University of Ari-zona. His research interests include systems support for mobile and pervasive computing. He focuses in particular on the challenges of creating deployable mobile and ubiquitous computing systems that can be used and evaluated “in the wild.” He’s an associate editor of IEEE Pervasive Computing. Contact him at the Computing Dept., InfoLab 21, South Dr., Lancaster Univ., Lancaster, LA1 4WA UK; [email protected].

Eyal de Lara is an assistant professor in the Department of Computer Sci-ence at the University of Toronto. His research interests include distributed, mobile, and ubiquitous computing. His long-term research goal is to provide ubiquitous access to information and services independent of location or the capabilities of specifi c devices. He received his PhD in electrical and computer engineering from Rice University. Contact him at [email protected].

Mirjana Spasojevic is a senior design researcher at the Yahoo! Mobile busi-ness unit. Her research interests include designing and deploying mobile and ubiquitous computing systems and human factors. She has been studying how and why people use camera phones, documenting current practices, and exploring opportunities for new technologies. She received her PhD in computer science from Pennsylvania State University. Contact her at [email protected].

William Griswold is a professor in the Department of Computer Science and Engineering at the University of California, San Diego. His research interests include ubiquitous computing, educational technology, software evolution and design, and aspect-oriented programming. He received his PhD in com-puter science from the University of Washington. He’s a member of the IEEE Computer Society and ACM. Contact him at [email protected].

JULY–SEPTEMBER 2006 PERVASIVE computing 23

privacy derived from their experience deploying a social networking appli-cation at a research conference. Their application strives to foster interaction between conference attendees by using public displays to visualize the social net-work that relates them. The application derives interconnected social clusters from the attendees’ publication record, obtained from the DBLP (DataBase Sys-tems and Logic Programming) biblio-graphical Web site, by looking at com-monalities in coauthors and citations. Even though their system relies only on information that is freely available on the Web, the authors note that its display in a public setting in overt relationship to nearby users raised privacy concerns. For example, some attendees with a limited publication history were disappointed by the small number of papers and coau-thors in their social network.

T here’s no doubt that resear-chers should think carefully before undertaking a major ubicomp system deploy-

ment. Deployments are often expen-sive and messy, can involve a lot of effort that isn’t really characterized as research, and might entail administra-tive concerns such as seeking ethical approval or planning consent. More-over, deployments typically require a sustained commitment to keep them operational and, in the case of public deployments, expose researchers to public scrutiny.

However, as Mark Weiser envisioned it, ubicomp is about computing “disap-pearing into the fabric” of everyday life. And the only way to effectively study how people interact with such “embed-ded” systems is to deploy, observe, and learn.

REFERENCES 1. J. Kjeldskov et al., “Is It Worth the Hassle?

Exploring the Added Value of Evaluating the Usability of Context-Aware Mobile Systems in the Field,” Proc. 6th Int’l Mobile HCI 2004 Conf., Springer, 2004, pp. 61–73.

2. R. Want et al., “An Overview of the Parctab Ubiquitous Computing Experi-ment,” IEEE Personal Communications, vol. 2, no. 6, 1995, pp. 28–43.

3. R. Want et al., “The Active Badge Loca-tion System,” ACM Trans. Information Systems, vol. 10, no. 1, 1992, pp. 91–102.

4. G. Abowd, “Classroom 2000: An Experi-ment with the Instrumentation of a Living Educational Environment,” IBM Systems J., vol. 38, no. 4, 1999, pp. 508–530.

For more information on this or any other com-puting topic, please visit our Digital Library at www.computer.org/publications/dlib.