Information Systems for Enhancing Social Experience ...

ISSN 03617688, Programming and Computer Software, 2012, Vol. 38, No. 6, pp. 281–293. © Pleiades Publishing, Ltd., 2012. Original Russian Text © Michael A.H. Fried, Anna Fensel, Federico M. Facca, Dieter Fensel, 2012, published in Programmirovanie, 2012, Vol. 38, No. 6.

Information Systems for Enhancing Social Experience in a Conference Context: A Sensor based Approach Michael A. H. Frieda, Anna Fensela, Federico M. Faccaa, and Dieter Fensela a

STI Innsbruck 6020 Innsbruck, Austria, Technikerstrasse 21a CREA TENET 38123 Trento, Italy, Via alia Cascata 56/D Povo email: [email protected], [email protected], federico.facca@createnet.org, [email protected] b

Received November 15, 2011

Abstract—We combine Auto Identification and Web technologies in an extensible onsite event support sys tem for enhancing experience of conference organizers as well as participants. Our system enables users to authenticate themselves using RFID badges and to access interactive, Webbased services via a touchscreen PC. The developed services aim at supporting social interactions of participants, and thus validate the prom ising usage directions of the combination of offline social networks and the online social Web. The main hypothesis of our work is that: “A sensor based information system can truly enhance social conference expe rience.” Technically, we have investigated employment of Web 2.0 technologies in social, sensor and mobile technologies enabled systems at conferences and events. This article gives an overview of the overall system and its evaluation via a user survey and usage log data collected during the Extended Semantic Web Confer ence (ESWC) 2010 and similar international conferences, altogether with several hundred participants. DOI: 10.1134/S0361768811020010

1. INTRODUCTION Over the last century human communication has evolved far beyond a scope imaginable by our prede cessors. Especially the digital advent with the world wide penetration through Internet connectivity, which ultimately lead to the ubiquity of communication devices that enable us to be theoretically connected to everyone one the planet 24/7, has significantly changed all our lives. However, despite the recent technological possibilities, humans always shared the need for communication. This behavior still holds true today, where social networking is a major aspect of conferences and simi lar events most participants benefit from. This net working includes activities like the interest in (scien tific) self comparison and the necessity to connect with other conference participants in order to gain new insights on various topics. Empowered by becom ing more and more affordable devices and sensors, ranging from regular TV sets, touchscreenenabled kiosks, to tablet PCs designated for the mass consumer market, RFID (RadioFrequency IDentification) and the ubiquitous smartphones, social networking is becoming supported by modern technology in a phys ical, onsite and locationaware manner, thus not only virtually as till the recent times. While actively coorganizing several annual con ferences on “Semantic Technologies and Future Internet we are convinced that innovative IT systems The article is published in the original.

have big potential to substantially support conference participants' networking activities onsite. For this purpose, we have developed and are offering easily accessible services to enhance the social experience during such events. The services are encapsulated in an extensible sys tem utilizing Web 2.0 technologies as well as sensors, which rely on automatic identification and data cap ture (AutoID) techniques, such as for example RFID and bar codes (2 dimensional QR codes) for authenti cation of users. These AutoID techniques are nowa days mostly used in supply chain management [11] and the logistics sector for tracking of assets and inven tory purposes, but are emerging as foundation of the Internet of Things [6, 17] and are starting to become ubiquitous in our everyday lives. Google1 recently started it’s “Wallet” service in the USA, which tries to replace the physical wallet with Near Field Communi cation (NFC, or also referred to as mobile RFID) enabled smartphones that can be used for all everyday monetary transactions. Our system and services have been deployed and partly evaluated in particular at the following confer ences: • European Semantic Web Conference (ESWC) 2009. • European Semantic Technology Conference (ESTC) 2009. 1 http://www.google.com/wallet/

281

282

FRIED et al.

Experiments Overview ESWC 2009 Part. Countries Users Services

approx. 300 approx. 40 approx. 80 Web Comparator VCard Exchange My Talks

ESTC 2009 approx. 100 approx. 10 approx. 20

ESWC 2010 Part. Countries Users Services

FIS 2010

312 (53 female) 112 (16 female) 46 20 120 11 Web Comparator Web Comparator VCard Exchange VCard Exchange Res. Impact Eval. Res. Impact Eval. ESTC 2010

Part. Countries Users Services

approx. 140 approx. 10 approx. 10 Web Comparator VCard Exchange Res. Impact Eval.

2. SYSTEM ARCHITECTURE AND IMPLEMENTATION

• Extended Semantic Web Conference (ESWC) 2010. • Future Internet Symposium (FIS) 2010. • European Semantic Technology Conference (ESTC) 2010. During the ESWC 2010 and FIS 2010 all user inter actions were logged to evaluate the system’s overall performance and the performance of different ser vices. Additionally over 30 people participated in a survey containing questions about the system’s usabil GWT client

get next particpant

GW

T

r we ser

get participan information

send id Framework user DB service 2

Our conference support system centers around a piece of software that enables conference organizers to deploy different services and conference participants to access them through traditional browsers and mobile clients. The authentication may potentially occur through different devices and technologies, such as Near field Communication, bar codes, and RFID tags. Moreover the user interface is designed to be intuitive and easy to use without requiring explicit instructions. Figure 1 presents the architecture of our system for onsite conference services. At the core of the archi tecture, there is a flexible authentication mechanism that allows users to authenticate themselves by using an RFID tag. Devices (e.g. such and RFID reader or mobile phones) send the authentication token to the server which looks up the user information in the underlying database. On successful login, users are provided with an overview page (Fig. 2) showing all available services. For later deployments the user database has been 2

send participant information

service 1

ity, service usefulness and user acceptance of different AutoID techniques. The collected data provides deep insights into the success of the project and indicates directions for the future of our research. The overall outcomes and contributions of the project include the evaluation of acceptance and use fulness of the services, identifying the social practices and types of services people prefer, AutoID/RFID acceptance, as well as techniques for promoting such a system during a conference and optimization of the interface and its adaptation to the requirements of a touchscreen or mobile device. The next section of this article discusses the overall architecture of the system, followed by a description of the developed showcase services in Section. Section describes the project’s history. Results form a detailed evaluation based on usage data collected during the ESWC 2010 can be found in Section. An outlook on Related and Future Work as well as Conclusions in Sections—finalize the article.

service 3

Fig. 1. Architectural overview.

integrated with the STI International Community website. This new platform aimed at becoming a social network for STI members and the Semantic Web com munity. Core features included browsing the commu nity, accessing the latest news from members on their projects and events, seeing showcases of semantic technologies, accessing public resources about seman tic technologies and semantic data and view related media created by STI and its members. Conference registrees were registered automatically to the com munity pages. They can could their profile informa 2 http://community.sti2.org/

PROGRAMMING AND COMPUTER SOFTWARE

Vol. 38

No. 6

2012

INFORMATION SYSTEMS FOR ENHANCING SOCIAL EXPERIENCE

283

Fig. 2. Service overview page.

tion and access the different Community subsystems: Agora (blogs), Emporium (marketplace), Kudos (challenges) and Minerva (wiki). To achieve the basic requirement of flexibility, the system is implemented using the Google Web Toolkit (GWT), a development toolkit for building complex Web 2.0 applications. GWT code is written in Java and later compiled to client side JavaScript that smoothly runs within most modern browser. Compatibility issues between different browsers as well as client server communication via GWT/RPC are handled automatically. However, services themselves do not have to be necessarily created with GWT. Our system can inte grate services written in any language or technology, such as PHP or Adobe’s Flash, that is able to be exe cuted within a browser. GWT based services will be added as modules, whereas others can be added with little programming effort (e.g. as a simple HTML page, external websites via iframe HTML tag, a HTML page with Flash embedded etc.). To enable interaction with the services by means of different devices, we adopted the ModelViewCon troller pattern: the conference GWT server exposes the system to the clients and handles the authentication mechanism. Each service exposes and accesses data through either a RESTful or GWTRPC interface. The client (browser) requests data either directly from the conference framework server, external websites using JSON with padding or external websites using the conference server as proxy (due to browser’s same origin policy). The retrieved data is then rendered on the client according to the defined service view. All user interactions are logged centrally on the server. At the moment we mostly concentrated our efforts on the touchscreen kiosk based version of the system and implemented a prototype using RFID for authen tication, but first concepts have been tested also for PROGRAMMING AND COMPUTER SOFTWARE

mobile devices (see http://iot.sti2.at). Section dis cusses our plans for mobile devices support. Hardware wise two independent passive RFID solutions (apart from the first installment at ESWC 2009 which featured an active solution) served as tech nical base for authentication: Until end of 2010, RFID hard and software was provided by RFIT Solutions located in Graz, Austria. Their RFID Lan Reader communicated to a PC using the Ethernet port combined with the YouR OPEN Software suite, a middleware solution which enables various kinds of AutoID devices to be accessed through a .net environment. The data is exposed through automatically generated standard WSDL Web Services, in order to be read into other systems. In general it is possible to store data directly on the RFID tags themselves, but due to memory limitations and a longer, more error prone reading process, the confer ence support system only uses the globally unique tag identifier and associates it with participant data. How ever, the Lan Reader turned out to offer too much unused functionality3 and overhead4 within the scope of this project. As a result, a new, more light weight RFID infrastructure was introduced. For the latest deployment, the authentication mechanism was switched to an OmniKey passive RFID reader. As opposed to the Lan Reader which was very big, powered by an external power supply and communicated via standard LAN network, the Omni Key reader is small, powered by and communicates over USB. The tags used in previous deployments were compatible with the OmniKey reader, which eased up the migration to the new piece of hardware. 3 Things like 3 freely programmable buttons and a small display to

show messages.

4 The need of a router in order to connect the reader to the touch

Vol. 38

screen PC together with an additional cable for the Internet connection. No. 6

2012

284

FRIED et al.

Fig. 3. RFIT Lan Reader (left) OmniKey RFID reader (middle) passive RFID tag (right).

The system’s touchscreen PC is shown in Fig. 4. 3. CONFERENCE SERVICES Conference participants have several typical social needs, such as contact information exchange, people and publication comparison, learning new relevant things, etc. Alongside the framework, as a proof of concept and to showcase the onsite conference sup port system, various service ideas have been devel oped. The presented infrastructure and services aim to address the different social needs of conference partic ipants. Not all of these services have been deployed at each conference. Table5 gives an overview of which services have been deployed where. The emphasis of this article is mainly to evaluate the usefulness and acceptance of the kiosk system itself and RFID tech nology. The service performance was evaluated, but these findings are secondary, since implemented ser vices so far are only meant to be a proof of concept for the implemented Web based conference support sys tem. 5 Some

values in the table are approximations due to missing or corrupted data.

Fig. 4. ESTC 2009 touchscreen PC and @RFID reader setup.

VCard Exchange service is inspired by the extremely common real world practice of business card exchange. The VCard application enables multi ple people to exchange their contact information at once. When logged in, the participants who want to exchange VCards can achieve this by simply wiping their RFID tags over the reader. The user information, which has been provided at the registration for the conference, is serialized to a VCard 2.1 [8, 4] compli ant format. This method for exchanging contacts was chosen, because it is very compact and compatible with most modern email clients and phones. On con firmation, VCards of all participating people are exchanged via email. Example VCard BEGIN:VCARD VERSION:2.1 FN:Michael Fried N:Fried;Michael;;; ADR:;;Tstrasse;Innsbruck;;A-6020;AUSTRIA TEL;WORK:0043987654321 TEL;FAX:004398765432100 EMAIL;INTERNET:michael.friedsti2.at TITLE:Researcher ORG:STI Innsbruck URL:http://www.sti2.at END:VCARD

Web Comparator showcases the boundaries of the current Web. It uses the names of up to four partici pants to lookup personal and professional informa tion on the Web and present the result to the partici pants. In particular we distinct three different sources of information: Web pages (Google, Yahoo), images (Google Images, Yahoo Images, Flickr) as well as sci entific publications (Google Scholar, CiteseerX). In this way participants jointly authenticated on the system can compare their respective presence on the Web. In the first versions of the system the comparison was summarized through an index, called “Web points that vaguely related to the number of results found. The missing transparency of how this index was calcu lated led to confusion amongst some users. Many of them did not intuitively realize the correlation between number of results and "Web Points” and were asking about their meaning. This flaw was eliminated by substituting them with the actual number of results found.


Vol. 38

No. 6

2012


The service shows that in the current Web it is either very easy or very hard to retrieve meaningful informa tion with just having a person’s name. People with common names often tend to disappear in the “white noise” of results. On the other hand people who have more exotic names and are actively present on the Web in social networks and such, are easy to find and some times very detailed personal profiles could be gener ated with just one simple search. Most computer sci entists should be well aware of their online presence and footprints they leave. It can be mainly seen as a fun application for people who do not know which variety of personal information is publicly available on the Web. In future versions the Web Comparator will experi ment with the use of semantic technologies (e.g. semantic search, consideration of Linked Open Data and inclusion of such into results, FOAF profiles, etc.) to reduce search result noise and compare traditional search results with semantically enhanced ones. Research Impact Evaluation is an outcome of the LiquidPub6 project. The project proposes a paradigm shift in the way scientific knowledge is created, dis seminated, evaluated and maintained. The shift is enabled by the notion of liquid publications, which are evolutionary, collaborative, and composable scientific contributions. In our terminal we integrate the ResEval [12] tool for evaluating research contributions and people by using citationbased metrics. The appli cation uses data of Google Scholar for calculating the metrics and is accessible via an online search form or RESTful Web service. Currently the information available contains Hindex [7], Gindex [5], N (number of publications) and C (number of citations) of an author and a list of all the author’s publications. Furthermore, for contri butions the citation count can be accessed. We plan to extend this service by including further functionality offered by the LiquidPub project, such as Research Network Comparison and Research Community Comparison. My Talks service enables participants to look at the schedule of the conference, decide which talks they want to attend and access their personal schedule on the terminal. An important point is the implementation of new services and improving existing ones based upon com ments collected in the user survey. Ideas for future ser vices include the following: Terminal Interaction Graph service visualizes the user interactions in a graph form using data provided by the semantic log. The graph shows the strength of connections (number of joined activities) between users and, when applicable, the Facebook profile pic ture to facilitate identification. A participant can 6 http://liquidpub.org/


285

review all previous interactions and send a Facebook friend request to people whom he or she has been jointly logged on with. Additionally publication data retrieved via DBLP is shown. Semantic Games is an entertainment service that can be used to create semantic content, a process which often cannot be solved automatically but requires certain human effort [15]. Programmed in Flash and accessible within the browser, it is very easy to integrate them into the terminal. The first one to be playable is SeaFish (SEmantic Annotation FISHing), a single or two player game used for image annotation. The game shows a reference image to its players who have to “catch” related images floating around on the screen. 4. HISTORY OF THE PROJECT Started in spring of 2009, as a small demo for the ESWC 2009, the application evolved and has since been deployed at the following events: 1. ESWC 2009: The first version of the system consisted only of the Web Comparator service presented on a non interac tive TV screen. Back then the underlying active RFID infrastructure for our system was supplied by the LSS [16] experiment, which also took place at conference. LSS supports and guides social networking activities between researchers at conferences and similar events by integrating data and technologies from the Seman tic Web, online social networks and an RFID tracking platform. Participants were supplied with active RFID tags and their movements and interactions tracked on site. Face to face interactions and social interaction graphs were visible on some screens within the confer ence area. According to the statistical information provided in [16], 305 people attended the conference, out of which 187 collected an RFID badge. 139 of the participants created an account on the LSS applica tion site and were therefore able to use our system as well. We used an active tag placed below the TV to rec ognize people near the device (1 to 5 meters) and start the Web Comparator service automatically. Due to the static presentation and the fact that the system often picked up signals from too many adjacent participants to guarantee proper authentication, the first experiment was not received exceptionally well by the audience. 2. ESTC 2009: Considering lessons learned from the first deploy ment, the system was upgraded to feature an interac tive touchscreen and an authentication mechanism based on passive RFID technology, which is a cheaper solution and suits the requirement of our scenario bet ter due to a more limited range. Instead of being rec ognized when approaching the terminal, participants had to wipe their uniquely numbered tag over a reader

Vol. 38

No. 6

2012

286

FRIED et al.

within a distance of 20 centimeters. The ESTC is a 2 days conference attended by around 100 people, with a stronger focus on business audience rather than researchers. At this conference we realized that the system’s promotion could be improved to draw more attention to the kiosk and its functionality. 3. ESWC 2010: After implementing new services, we decided to monitor activities on the terminal and to conduct a user survey for evaluating the system and service acceptance. Due to the size of the conference we expected to collect comprehensive data in order to fur ther improve the experience and get new ideas for future development. The data received from the sys tem logs as well as the survey will be evaluated in Sec tion of this article. The findings helped us to find weak spots in the ser vices and define a future direction for the project. 4. FIS 2010: No new services were introduced for the confer ence. The focus was to fix minor issues that were pointed out in course of the ESWC 2010. 5. ESTC 2010: Basically the same setup as for FIS with some minor changes in VCard serialization. 5. EVALUATION The data analyzed in this section was collected dur ing ESWC 2010. We wanted to identify how frequently people use the system and which services are popular. In 2010, 312 people were registered for the ESWC out of which 83% were male. RFID tags were produced for all of them, but some participants who registered on site did not pick up their tags. In total 120 people used the system which is approximately 38% of all attend ees. The audience was international, yet 89% came form Europe, 21% from Germany, 12% from the UK and 7% from Austria. Remote usage logging from the client side was implemented by sending log messages to the server when a person either logged onto the sys tem, chose a service or was added to one. The infor mation contains the kind of interaction, unique ID of the user, as well as a time stamp. Participants had to agree that their personal data could be used, via a check box integrated in the online conference registra tion form, in order to receive an RFID tag. Only geo graphic data as well as gender information were kept for performing various statistical analysis. Names, e mail addresses and all other personal information was deleted after the conference. The anonymization of data guarantees that a particular data set cannot be tracked back to one single participant. The system was promoted during the welcome speech and by an explanatory poster behind the terminal, which was

prominently placed next to the conference registration desk, which was the single spot where every participant passed by at least once in order to receive a conference badge. Additionally one person was present most of the time in order to explain the system and answer questions risen by users. In addition to the collection of usage logs, a feed back questionnaire was placed next to the terminal. Thirtyseven attendees participated in the survey, i.e. about 11% of the conference participants and over 30% of the people who actually used the system. Collected over the course of five full days, the logs unveil a large variety of facts. With a total of 587 logged interactions, Fig. 6 indicates that most people used the terminal up to five times. One participant had 39 inter actions, but all users that interacted with the system over 12 times can be considered statistical outliers. Over half of the people used the kiosk on exactly one out of the five days, whereas three registrees only logged in without selecting a service. May 31st was the busiest day, June 3rd the weakest. Male as well as well as female usage show a peak on the second day. Distribution of the usage times are shown in Fig. 7. This graph clearly shows more interactions on the first days of the event and almost a flat interaction curve during the last day. Over the course of one day the graph shows different peak times. Unsurprisingly, the terminal was used more often during lunch and coffee breaks, with increased usage around 11 am, 1 pm and 4:30 pm. Although it was not the busiest day, June lst showed the overall highest kiosk usage between 3:30 pm and 4:30 pm, whereas no one used the system on the last day after 5 pm. A correlation between terminal usage and general attendance level cannot be proven, but our observa tions of the conference venue during the deployment indicate that more people were present on the second day than on the last. Results shown in Fig. 8 provide insight on how frequently people used the different services: • Research Impact Evaluation drew the least user attraction (slightly over 60 users), but had fairly many interactions. This means that everyone using the ser vice used it three times on average. • VCard Exchange had the smallest number of interactions as well as reuse ratio. On average each of the service’s users exchanged a VCard twice. • Web Comparator was the most popular service by user base (90 users) as well as interactions (over 200). This distribution unveils that there is a clear gap between the two, very similar, comparison services (Res. Impact Eval. and Web Comparator). Despite the assumption that IT people are well aware of their online presence, the Web Comparator drew the biggest attraction on the conference. Surprisingly the VCard Exchange service did not perform as well as expected. This could result from the fact that some people rather


Vol. 38

No. 6

2012


287

Lat 80 60 boxes

40

1 2−5 5−10 10−25 25−50 >50 NA

20 0 −20 −40 −150 −100

−50

0 Long

50

100

150

Fig. 5. Participant countries overview World.

Nunber of participants

Times used

20 15 10 5 0

30

1 times 2 times 3 times 4 times 5 times 6 times 7 times 8 times 9 times 10 times 11 times 12 times 13 times 14 times 15 times 16 times 18 times 19 times 39 times only log on

20 10

Times used

0 Participants

70 60 50 40 30 20 10 0

40

ov er all

Nunber of participants

30 20 10 1 day

2 days

3 days

4 days only log on

ale m female

30 May

31 May

Number of days used

01 Jun

02 Jun

03 Jun

Date

Fig. 6. Terminal usage statistics. Overview of how many times/on which days users accessed the terminal.

distribute traditional business cards than electronic ones. Another theory is that in a closed and compara bly small community like the one represented in ESWC7 most people simply know each other already. If we compare our deployment to the results from the IntelliBadge project [3], we see various common alities in the data. Out of the 2188 IEEE Supercom puting 2002 (SC 2002) registrees 872 people could be 7 Many

European core researchers within the Semantic Web community frequently attend a limited number of annual con ferences. PROGRAMMING AND COMPUTER SOFTWARE

tracked as using the system. This results show a similar user/conference participants ratio to our experiment. Also the usage declined after a peak on the first day. The IntelliBadge kiosks were accessed 1771 times on site and 1370 times remotely. Every participant used the system about 4 times on average. Also the results of the LSS experiment [16] show a similar picture. Out of the 305 participants 187 people received an RFID tag and 139 used the system. When taking a look at the survey feedback evalua tion, the results provide deeper insight into service and

Vol. 38

No. 6

2012

May 30 May 31

20 15 10 5 0 20 15 10 5 0 20 15 10 5 0

June 3

20 15 10 5 0

June 1

FRIED et al.

June 2

288

20 15 10 5 0 09:00

11:00

13:00

15:00

17:00

19:00

Time Fig. 7. Usage times.

RFID acceptance. The questionnaire contained cer tain terminal specific questions which are depicted in Fig. 10. Answers could range from 1 to 7 (where 1 marks the best grade and 7 the worst). Additionally Fig. 9 contains an overview of RFID specific questions. Two free text fields offered the possibility to provide “General notes/issues” and comment on the question “What services would you like to see in the future?”. Some frequent answers hint at requirements desired by participants: 1. Integration with social networks. 2. Bulletin board service. 3. List of conference participants with photos to allow identification. 4. Recommendations based on relations. 5. One kiosk is not enough. 6. Access services over mobile phones.

The social needs addressed by the services include the interest in self comparison and the necessity to connect with other conference participants. Results collected with the usage logs and survey indicate that the services performed well and met the users' demands. An interesting fact is the gap between the grading and usage of the Web Comparator and VCard Exchange service that can be seen in Fig. 10. Whereas the latter had scored the best grade in the survey and has been elected the most useful service, it has not been used the most. The Web Comparator was graded least but turned out to be the most popular service. Reasons for this usage distribution can result from various factors like service responsiveness, attractive ness, the fun factor, etc. This article does not focus on explaining reasons for the discovered distributions, but rather provides an overview of general usage which has to be investigated further in future setups by adapting


Vol. 38

No. 6

2012

INFORMATION SYSTEMS FOR ENHANCING SOCIAL EXPERIENCE Number of users

289

Number of users

80

200

60

150

40

100

20

50

0

0 Res. Impact. Eval.VCard Exchange Web Comparator

Res. Impact. Eval.VCard Exchange Web Comparator

Service Number of users 3.0 2.5 2.0 1.5 1.0 0.5 0

Service

Res. Impact. Eval.VCard Exchange Web Comparator

Service Fig. 8. Service usage statistics. Overall number of service users and service interactions.

Did you come in contact with RFID before?

Do you have issues redarding RFID?

answer No

answer No

Yes, being tracked

Yes

Yes, health issues Yes, Other

Would you prefer barcodes over RFID? answer Barcodes Barcodes or QRCodes I don't care No QRCodes

Fig. 9. Overview and results of RFID specific survey questions.

the feedback survey and trying different designs and service layouts. Although most of the people came in contact with RFID before the conference, there is still skepticism towards RFID technology and users tend to prefer bar codes. We will address these concerns in future deployments of the system (e.g. information that no PROGRAMMING AND COMPUTER SOFTWARE

tracking is performed, introduction of bar codes, etc.). Nevertheless, the use of sensor systems for authentica tion has proven so far to be a secure and easy to use solution. The survey shows that there are still improvements to make regarding terminal usability, interface design and overall experience. Yet, all in all, the use of RFID

Vol. 38

No. 6

2012

290

FRIED et al. Question Is using RFID to offer services at conferences useful? Grade your overall experience with the experiment. Grade the Web Comparator service. Grade the VCard Exchange service. Grade the user interface's dessign. Grade the terminal usability. Grade the Research impact Evalution service.

7

6 5 4 3 2 Mean user feedback value

1

Which is the most useful service?

service Research Impact Evaluation VCard Exchenge Web Comparator

Fig. 10. Overview and results of terminal specific survey questions.

at conferences to offer services has been graded useful by the majority of participants. 6. RELATED WORK As a real world application, the proposed system touches multiple different fields of research: Web Engineering (exposing the system via a browser), AutoID Systems (use of RFID or QR Code for authentication) and Social Sciences (usability and acceptance studies). The core idea of the project, i.e. to add value to conferences through technology means, has been occurring in research. For example the adoption of RFID for authentication is quite established in litera ture [2]. In their work, J. Bravo et al. focus on touch less interaction with a static kiosk. In comparison, our 1 solution emphasizes haptic user interaction via a touchscreen and provides an interactive and simple way to expose and access services. Our services pro mote social networking and interaction among con ference participants. Ubiquitous community assistance (UbiCoAssist) and POLYPHONET [13] address as well the notion of services for social networking during conferences. These tools were showcased during UbiComp2005 and UbiComp2006 and contain a set of specialized services to point out social network relations between conference participants.

During the first demonstration of our system at ESWC 2009, we strongly interwove it with the ESWC Live Social Semantic (LSS) experiment [16], which was first described in [1]. LSS tracked conference par ticipants using active RFID technology and we built on top of their infrastructure to support identification at our kiosk. A similar approach for tracking partici pants was experimented by the IntelliBadge [3] project at the IEEE Supercomputing conference in 2002. Data collected by IntelliBadge and LSS provides us reference values for evaluating the performance of our system as discussed in Section. Some commercial solutions have been developed as well, such as Spotme8. During conferences every participant receives a proprietary handheld device that enables them to access services and exchange data with other people onsite. Spotme integrates social networking, audience response, and participant man agement tools. In comparison to such systems, our solution adopts standard Web technologies that make our system usable on different devices and simplify its development and integration: it is comparatively easy to develop new services and integrate existing ones. The adoption of RFID for authentication grants a certain degree of security, since tags cannot be faked or reproduced trivially. RFID is slowly becoming ubiqui tous in our every day life and several privacy thread and 8 http://www.spotme.com/


Vol. 38

No. 6

2012

INFORMATION SYSTEMS FOR ENHANCING SOCIAL EXPERIENCE interaction

hass Timestamp

is Subclass Of

xsd:datetime user

performs

logon

291

xsd:string

logoff

has Name

calls

service interaction

is Added To

service

Fig. 11. Semantic interaction ontology.

issues [10] (e.g. the ability to track people) have been raised on this technology. Inspired by existing research [9] we dedicated a part of our survey to the users' acceptance of AutoID systems. A very recent paper [14] takes the idea of support ing social conference experience one step further by completely virtualizing the whole event. The author describes how an internal IBM event (Academy Gen eral Meeting) that was formerly a facetoface meeting has been held in the virtual online 3D world “Second Life”. This platform owned and run by Linden Research, Inc9 triggered a big buzz in recent years, when companies started to feel the need to be present in this community. With over one million users, the system it is still a in operation today, but mainstream media reports have significantly dropped in the last 2 years and the hype seems to be over. This measure was part of an initiative by IBM to save costs in the light of recent economic crisis'. Thus in 2008 and 2009 these conferences were held in a private section of “Second Life”. The outcome of the two experiments showed that when regarding current stateoftheart, it is very unlikely that conferences in a completely virtual space can replace facetoface meetings. Common issues arise when navigating through the conference space: Identification (to whom the user is speaking), naviga tion (how to get around), conversation scope (setup of private chat). 7. FUTURE WORK A key finding of the evaluation is that a significant amount of people, over one third of all participants, used the system, but the provided infrastructure is not sufficient for large conference setups, i.e. one touch screen terminal is not enough, since it is often occu pied and conference venues can be very large, which means that participants could have to walk an ample distance in order to simply exchange a VCard for example. So spatial constraints, even if the touch screen is placed in a highly visible spot, could discour age people from using the system. Ultimately a shift to 9 http://secondlife.com/


mobile devices appears to be necessary. There are strong indications that mobile RFID via NFC will be featured in many future phones, as Google has already integrated the technology in it’s Nexus S branded phone and as Apple is allegedly looking into this tech nology for future versions of the iPhone. Some exper iments based on Google’s Android smartphone OS platform are already ongoing. Furthermore a focus on deeper integration of social networks seems crucial. The integration of social net works within system provides a strong social notion for future service development. Apart from the user and community management integration, semantic tech nologies could play a key role in the following compo nents and approaches: • A user will be able to connect a Facebook account to our system and use Facebook’s functional ity by means of different future services (e.g. people recommendation based on friends, automatic posting to the Facebook wall when using a service, adding peo ple as friends when exchanging information, etc.). • All user interactions will be stored in a semantic log, within an RDF triple store, according to the inter actions ontology shown in Fig. 11. This information can then be reused by new services (e.g. Terminal Interaction Graph service, people recommendation, evaluation). The log unveils social connections estab lished by using the terminal and enables onthefly usage statistics. The interaction ontology was kept simple and generic in order to guarantee flexibility for new services. • Data used within implemented and planned ser vices originates from various sources such as the Goo gle Search REST API and the Linked Open Data (e.g. FOAF profiles, DBLP publications, information from DBPedia, etc.) cloud. Also on the other side, services of the system could potentially serve as generators of semantic annotations. Moreover, work on this system and especially the outcome of the user survey has risen our interest in data dissemiation and retrieval. During our research we have realized that the online world is becoming very fragmented, with a vast amount online channels emerging and quickly changing over time. For exam

Vol. 38

No. 6

2012

292

FRIED et al.

ple small and medium sized businesses can easily loose track of stateoftheart and miss important future trends. There is a very huge shift from a mainly static to a very participatory Web, with the smart interlinking of data shifting the Web of Documents towrds a Web of Data. In order to provide a coherent solution in the constantly changing market we are currently working on a solution for multi channel dissemination, feed back aggregation and analysis. 8. CONCLUSIONS Extending conventional offline social networking activities, the presented system has been generally well received at the conferences. Existing services within the system have proven a valid and meaningful integra tion of very heterogeneous technologies, including RFID and Web 2.0. Future services will continue the idea of supporting ubiquitous social interaction of conference participants and foster integration of Mobile, Semantic and Social Web into the system. The system is designed to be easily extended with multipurpose services, not only addressing the (com puter) science community, but a far more general audience. One of the aims of the project has been to test and evaluate the use and acceptance of AutoID technol ogy in conference scenarios. Here a key finding is that relatively many people have concerns with the use of RFID technology that need to be carefully addressed when designing and deploying such a kind of systems. As the evaluation shows, even more social, recom mendation and dissemination features are desired, and new services such as talks/agenda recommenda tion and social network integration are important for the future success of the project. Altogether the sys tem’s ultimate goal is to grow into a social interaction hub for conferences and similar venues, especially the ones with a large number of participants, where it has been most demanded. In detail we draw the following conclusions about our system’s contribution to the field of ICT research: • How can social onsite networking be supported and extended by a Web based system that makes use of Web 2.0 techniques? —Contribution: A prototypical framework, created with GWT using AJAX techniques, to enable the exposition of multipurpose services within a browser. Deployment of the system on a touchscreen PC with an RFID enabled mechanism for user authentication. —Conclusion: Although the system was used and well received, the extent to which the system in the end supported social onsite networking was quite limited. The emphasis of this thesis lies on the framework, but the things that naturally drive the system performance are the offered services. The infrastructure is now there and new services have to be implemented in upcoming steps.

• How can such a system be promoted during an event? —Contribution: Different means of promotion, ranging from conveniently locating the kiosk, to dis tributing handouts, etc. were tried out in course of the events. —Conclusion: It is very hard to exactly point at which factors really influenced the system’s perfor mance. Similarly the impact of the promotion is hard to measure. However some general points, like the location of the system and someone onlocation to provide further information to interested participants were crucial means to attract users. The unstructured trial and error approach to this problem taken by the project, turns out to be insufficient in order to provide wellfounded suggestions on how to promote similar systems. • Is such a system useful in a real world scenario and accepted by users? —Contribution: The evaluation performed in within this thesis, user data collected during the deployments as well as the user survey provide insights into usefulness and user acceptance. —Conclusion: According to the survey, most par ticipants found the system and implemented services rather useful. There is much space for improvement especially on the available service side. The system itself was well accepted and, especially at large events, it was used widely. However there is much space for future improvements. These findings resulting from the evaluation and based upon lessons learned, will govern the future direction of the project. • Do people accept RFID technology or are there any concerns regarding this or other AutoID systems? —Contribution: A part of the user survey was dedi cated to AutoID systems and especially RFID accep tance by the system’s users. —Conclusion: Despite the fact that most partici pants came in contact with RFID before and seemed to be well informed about the technology, many par ticipants showed prejudices against it (cf. Fig. 9). The biggest issue seemed to be the anxiety of being tracked anywhere on the conference site. Not only was this impossible with the passive RFID tags used in most experiments, but rather not even desired within the scope of this project. A minority of participants really vouched for RFID technology and half of them would rather see different technological means of authenti cation. However RFID tags and readers proved to be secure and easy to use, which are two definite plus points for the technology. Apart from this, the mone tary factor is a strong argument against the technology and its economically weakest point. Depended on the scenario RFID can add value to certain systems. In the framework presented in this thesis, however, it can be easily substituted.


Vol. 38

No. 6

2012


REFERENCES 1. Alani, H., Szomszor, M., Cattuto, C., Van den Broeck, W., Correndo, G., and Barrat, A., Live Social Semantics, The Semantic WebISWC 2009, 2009, vol. 5823, pp. 698–714. 2. Bravo, J., Herva’s, R., Sa’nchez, I., Chavira, G., and Nava, S., Visualization Services in a Conference Con text: An Approach by RFID Technology, Journal of Universal Computer Science, 2006, vol. 12, no. 3, pp. 270–283. 3. Cox, D., Kindratenko, V., and Pointer, D., Intelli Badge, Proc. of 1st International Workshop on Ubiqui tous Systems for Supporting Social Interaction and Face toFace Communication in Public Spaces, 5th Annual Conference on Ubiquitous Computing, Seattle, WA, 2003, pp. 41–47. 4. Dawson and F., Howes, T., vCard MIME Directory Profile, http://www.rfceditor.org/info/rfc2426 (1998). 5. Egghe, L., An Improvement of the hIndex: The g Index, ISSI Newsletter, 2006, vol. 2, no. 1, pp. 8–9. 6. Gershenfeld, N., Krikorian, R., and Cohen, D., The Internet of Things, Scientific American, 2004, vol. 291, pp. 76–81. 7. Hirsch, J.E., An Index to Quantify an Individual’s Sci entific Research Output, Proceedings of the National Academy of Sciences of the United States of America, 2005, vol. 102, no. 46. 8. Howes, T., Smith, M., and Dawson, F., A MIME Con tentType for Directory Information, http://www.rfc editor.org/info/rfc2425 (1998). 9. Juels, A., RFID Security and Privacy: a Research Sur vey, IEEE Journal on Selected Areas in Communications, 2006, vol. 24, no. 2, pp. 381–394.

10. Lee, H. and Kim, J., Privacy Threats and Issues in Mobile RFID, Proceedings of the First International Conference on Availability, Reliability and Security (ARES), 2006, pp. 5. 11. McFarlan, D., Sarma, S., Chirn, J.L., Wong, C.Y., and Ashton, K., Auto ID Systems and Intelligent Manufac turing Control, Engineering Applications of Artificial Intelligence, 2003, vol. 16, no. 4, pp. 365–376. 12. Muhammad, I., Marchese, M., Ragone, A., Birukou, A., Casati, F., Jara, L., and Juan, J., ResEval: An Open and Resourceoriented Research Impact Evaluation Tool, Ingegneria e Scienza dell’Informazi one, University of Trento, Technical Report. DISI10 016 (2010). 13. Nishimura, T., Matsuo, Y., Hope, T., Hamasaki, M., et al., Casual Interfaces for Ubiquitous Community Assistance, CSCW 2006 Workshop, Collaborating over Paper and Digital Documents, Banff, 2006. 14. Sempere, A., Architecture and Design for Virtual Con ferences: A Case Study, Future Internet, 2011, vol. 3, pp. 175–184. 15. Siorpaes, K. and Hepp, M., Games with a Purpose for the Semantic Web, Intelligent Systems, IEEE, 2008, vol. 23, issue 3, pp. 50–60. 16. Szomszor, M., Cattuto, C., Van den Broeck, W., Barrat, A., and Alani, H., Semantics, Sensors, and the Social Web: The Live Social Semantics Experiments, 7th Extended Semantic Web Conference, Heraklion, 2010. 17. Welbourne, E., Battle, L., Cole, G., Gould, K., Rec tor, K., Raymer, S., Balazinska, M., and Borriello, G., Building the Internet of Things Using RFID: The RFID Ecosystem Experience, Internet Computing, IEEE, 2009, vol. 13, issue 3, pp. 48–55.

SPELL: 1. haptic


293

Vol. 38

No. 6

2012