CONTEXT-AWARE WEBLOG TO ENHANCE COMMUNICATION

AMONG PARTICIPANTS IN A CONFERENCE

Kosuke Numa

The Graduate University for Advanced Studies

Hideaki Takeda

National Institute of Informatics

Takuichi Nishimura, Yutaka Matsuo, Masahiro Hamasaki, Noriyuki Fujimura,

Keisuke Ishida, Tom Hope, Yoshiyuki Nakamura

National Institute of Advanced Industrial Science and Technology

Satoshi Fujiyoshi, Kazuya Sakamoto

Alpha Systems INC.

Hiroshi Nagata, Osamu Nakagawa, Eiji Shinbori

Dai Nippon Printing Co., Ltd.

Keywords:

Academic conference support, Weblog, communication support.

Abstract:

In this paper, we propose a Weblog system called ActionLog, which can associate Weblog entries to real world

contexts. The real world context is not only useful for Weblog authors themselves, but also beneﬁcial to

communication among people, because people with the same or a similar context can easily ﬁnd each other.

ActionLog collects user actions from both Web-based and other real-world systems. Weblog contents are

generated automatically from these user actions with related information. Users can browse and edit these

entries as their personal action records and in addition can publish them for experience sharing. We built a

practical system as an ActionLog application for conference support. Its main objective is to help people to

learn about presentations and to effectively meet other people in a limited amount of time and the least amount

of user commitment. There are three main features, i.e., integration of information in the real world and the

cyber world, personalization in both information capturing from users and in information providing to users,

and persistent assistance throughout the conference.

1 INTRODUCTION

Finding people with similar interests is an important

activity in event spaces, such as academic confer-

ences, conventions, and exhibitions. Thanks to quick

diffusion of the Internet, our ability to obtain infor-

mation and to know people is enhanced very much.

We can meet people regardless of time and distance

and we can obtain huge information instantly. On the

other hand, our life becomes more complicated than

one in the previous age. We should choose appropri-

ate ways among various Internet applications depend-

ing on circumstances. Furthermore we do not live on

the Internet all the time. Rather we still spend more

time in the real world. So we should switch to real-

world interaction from Internet interaction and vice

versa frequently.

In this paper, we propose a Weblog system called

ActionLog, which can associate Weblog entries to the

real world context. The real world context is not only

useful to Weblog authors, but also beneﬁcial for com-

munication among people because people with the

same or similar contexts can easily ﬁnd each other.

ActionLog is the concept for associating the real

world context to the Weblog. We are now building

two types of ActionLog applications. One is an appli-

cation for open environments in which GPS-equipped

mobile phones are used for capturing user context.

The other is an application for conference support.

We applied our system to two academic conferences.

In this paper, we mainly report the system applied

to UbiComp05, which is an academic conference for

ubiquitous computing.

The paper is organized as follows. In the next sec-

tion, we describe the concept of ActionLog. Then the

sections following it will explain the ActionLog sys-

tem for UbiComp05.

400

Numa K., Takeda H., Nishimura T., Matsuo Y., Hamasaki M., Fujimura N., Ishida K., Hope T., Nakamura Y., Fujiyoshi S., Sakamoto K., Nagata H.,

Nakagawa O. and Shinbori E. (2006).

CONTEXT-AWARE WEBLOG TO ENHANCE COMMUNICATION AMONG PARTICIPANTS IN A CONFERENCE.

In Proceedings of WEBIST 2006 - Second International Conference on Web Information Systems and Technologies - Internet Technology / Web

Interface and Applications, pages 400-407

DOI: 10.5220/0001253904000407

 SciTePress

2 BASIC IDEA

Weblog has been widely spread all over the world.

Many people use Weblog as a record of their dialy

life like what they eat, whom they meet, what they

buy and so on. Our system supports such authoring

by adding context.

We model users’ behaviors as actions with real

world contexts. Action means what they do and the

real world context is a set of information for the envi-

ronment in which they do so. The real world context

includes some physical situations like time and place,

subjective situations like intentional objects, and so-

cial situations like people nearby.

The system manages to capture actions and con-

texts by using sensors and cooperating with other sys-

tems. Some contexts, like time and place can be ob-

tained directly from sensors, but others should be in-

ferred from sensor information. For example, we can

obtain nearby people by integrating locational infor-

mation of multiple users.

Then the system generates an entry for each action

with sentences indicating the context of the action as

a draft. Furthermore, the system shows related infor-

mation, e.g., entries by other people with the same or

similar contexts. This is achieved by integrating and

calculating contexts.

The ActionLog usage scenario is as follows (Fig-

ure 1). The user walks around and does something

with a wearable device like a mobile phone. They no-

tify the device when they want to record an action.

After they go back home, they browse with a PC a list

of drafts, each of which corresponds to their actions,

and edits and publishes some of them.

Figure 1: Usage scenario of the ActionLog system.

3 ACTIONLOG FOR

CONFERENCES

In the previous section, we explained the ActionLog

in general. In this section, we focus on an ActionLog

to support academic conferences.

3.1 Supporting Academic

Conferences

Finding people with similar interests is an important

activity in event spaces, such as academic confer-

ences. Presenting information related to participants

is helpful for this purpose. However, because of a

time limitation and effort, it is almost impossible to

look through all the information. It is also difﬁcult to

write and publish one’s own information for sharing

with other participants.

In this research, we help people with the idea of Ac-

tionLog. There are three features in the proposed way

of assistance. One is that it offers integrated services

between the real world and the cyber world. The sys-

tem can gather and integrate users’ information, like

their actions both on site and on the web. Further-

more, users can access services either on site or from

the Web. The second feature is that it is fully per-

sonalized. It does not simply mean that information

can be selected according to users’ proﬁles but that

users can use the system as an assistant throughout

the conference, i.e., they can settle a schedule, specify

their acquaintances, and add their comments or jour-

nals with the system. The third feature is that it offers

persistent services. Users can start to use the system

before the conference and continue to use it after the

conference without loosing information.

We built a conference support system called the

Ubiquitous Community Support System (UbiCoSup)

and ActionLog for a Conference (ActionLog-C) is one

of the component systems of the UbiCoSup. Its main

objective is to help people to learn about presenta-

tions and to effectively meet other people in a limited

amount of time and with little user commitment. The

basic functions of the UbiCoSup are: (1) visualizing

researchers’ social network, (2) presenting informa-

tion about presentations and demonstrations, and (3)

sharing experiences by using automatic Weblog gen-

eration from users’ actions. The third one is realized

by ActionLog. The system collects information by

mining from the Web, by following users’ instruc-

tions, and by capturing users’ actions from systems.

Then it integrates and presents the information.

3.2 Action-oriented Experience

Sharing Support

The purpose of ActionLog-C is not only to sup-

port participants in composing their record of expe-

riences but also to share experiences among partici-

pants. Sharing opinions or impressions among partic-

ipants for an event serves an important role in charac-

terizing the event and in communicating about it with

other participants. We propose a way to share par-

ticipants’ experiences using ActionLog-C to achieve

CONTEXT-AWARE WEBLOG TO ENHANCE COMMUNICATION AMONG PARTICIPANTS IN A CONFERENCE

401

casual sharing of opinions among participants.

In this research, we assume an experience consists

of an objectively observed action and her subjective

thought at that time. To achieve experience sharing

using information systems, we propose a method to

capture user’s actions, to generate contents from these

actions, and sharing the contents among participants.

Based on this method, the assistant agent helps the

user to communicate with other participants.

3.3 Criteria to Design

Communication

To design communication by sharing action-based

contents, the following points should be considered.

What kinds of actions are shared? We especially

address event participants’ interactions with informa-

tion systems because they seem to be active and spon-

taneous actions for the system user. These interac-

tions are efﬁciently collectable by the system and are

easily understandable for participants.

When actions are shared? Meeting the other partic-

ipants is an especially important activity in an event

space. Action-based contents should be ready for

sharing immediately after the actions occurred to sup-

port participants in ﬁnding participants with identical

or similar interests. Previous works for sharing ex-

periential contents do not process the experience in

real time. PEPYS (Newman, 1991) and ComicDi-

ary (Sumi, 2002) are also automatic content genera-

tion systems that generate diary-style contents day-

by-day.

How actions are shared? It is important to reduce

users’ costs of capturing actions, creating contents,

and sharing contents. Wearable computers can record

all things the user sees and hears (Gemmell, 2002).

Video and audio recordings help users to review their

actions and share those experiences; nevertheless, us-

ing such large-scale equipment engenders high costs

and is difﬁcult to casually achieve. Sharing videos is

also a difﬁcult task for watchers of the contents. Text-

based and/or other easily understandable expressions

of experiences are needed.

With whom actions are shared? Sharing actions

engenders privacy problems. As for location-based

systems, some solutions, such as an anonymizing ap-

proach (Gruteser, 2003) or an obfuscation approach

(Duckham, 2005), are proposed. However, in event

spaces where participants wish to meet people, dis-

closure of the location and/or actions does not matter.

It is sufﬁcient for the user to select to publish or hide

each action using a one-click operation.

3.4 Basic Design for Experience

Sharing

To involve user’s subjective thoughts to the contents,

personal experience should be described by the user.

For example, we cannot tell what Mr. A is doing when

we ﬁnd the situation where he is in a restaurant, eat-

ing some food, and talking about something with Ms.

B. Mr. A might say he was having dinner, having

an important discussion, or just dating with Ms. B.

Therefore, we employ Weblogs as an infrastructure

for contents sharing. People can easily express their

thoughts or opinions with Weblog. In addition, We-

blog offers an infrastructural aid to manage and share

machine-readable contents (Karger, 2004).

The generated and published contents are aggre-

gated according to their context. For example,

extracting and listing the contents about a certain

presentation achieves a topic-centric view (Avesani,

2005) on the presentation. However, to communicate

with others, sharing contents is insufﬁcient. Allowing

users to reply to other contents can produce discus-

sions.

The architecture of ActionLog-C is shown in Fig-

ure 2. The left-hand side of the ﬁgure indicates the

action collection part of ActionLog-C. The system

captures the user’s actions from the other subsystems

of the UbiCoSup. We describe this part in the next

section. The center part indicates the core part of

ActionLog-C, which can gather and accumulate in-

formation from other subsystems and offer it to users

with different views. We explain the details of this

process in Section 5.

Figure 2: System architecture of the ActionLog system.

4 CAPTURING USERS’ ACTIONS

4.1 Overall

ActionLog-C collects the user’s action from other

UbiCoSup subsystems. UbiCoSup consists of three

WEBIST 2006 - WEB INTERFACES AND APPLICATIONS

402

Web-based and three on-site subsystems as follows:

• Web-based subsystems:

– Researchers’ social network system (part of

Polyphonet Conference)

– Scheduling support system (part of Polyphonet

Conference)

– Action-oriented Weblog system (ActionLog)

• On-site subsystems:

– Kiosk PC terminal using RFID card for user au-

thentication

– Photograph sharing and visualizing system

based on an omni-directional camera and RFID

card (Tabletop Community)

– Presentation annotating system based on mobile

telephone (Information Clip)

These subsystems are built for the purpose of sup-

porting participants in academic conferences. Re-

searchers’ information and their social relations are

useful for understanding the research community

and for contacting them. The scheduling sup-

port system helps participants to check presenta-

tions/demonstrations, by showing information about

them i.e., authors, titles, and abstracts. Users can add

a presentation to her schedule if she likes it. Par-

ticipants can access the Web-based subsystems and

view the social network using the kiosk PC settled

in the conference site. Participants can communi-

cate with others on the Tabletop Community, viewing

their photographs and their networks. The Informa-

tion Clip provides ways of checking presentations on

site, annotating them with a mobile telephone. The

user scans QR codes or RFID tags that are attached

to panels set close to poster presentations and demon-

strations using camera-enabled mobile telephones or

mobile telephones with RFID tag readers, then the

user can add comments on the presentations, and rate

them.

User’s action records on these subsystems are cap-

tured and integrated by ActionLog-C. We brieﬂy in-

troduce other UbiCoSup subsystems and deﬁne the

actions captured from them.

4.2 Polyphonet Conference

The Polyphonet Conference is a social network

browser and a conference scheduling system. Poly-

phonet obtains attendees’ relationships and research

topics from web sites. A user can ﬁnd what research

topic a particular researcher is doing or whom they

are working with. In the scheduling part, a user

can register interesting presentations (papers, demos,

and posters) and get recommended presentations and

other researchers.

Polyphonet has three methods to extract social net-

works among participants. The ﬁrst one is based on

Web mining techniques. A Web mining method is

applied to extract a social network among the par-

ticipants (Matsuo, 2003). Roughly speaking, our

approach is to measure the relevance of two nodes

based on the number of retrieved results obtained by

a search engine query, e.g., “Mark Weiser and Alan

Kay.”

The second one is based on a user’s operation on

the Web system, and the third one is captured from

users’ interaction on the conference site.

Two types of actions are captured. One is to add

a person to the acquaintance list (Action 1), and the

other is to add a presentation to someone’s own sched-

ule (Action 2).

4.3 Kiosk PC

We place kiosk PCs at the conference site. Partici-

pants can access Web systems. We distribute RFID

cards to participants. The kiosk uses RFID cards

for user authentication. Each kiosk has three RFID

card readers. A user can log into the Web system

by putting her card on one of card readers. In addi-

tion, participants can see social networks among them

when two or three participants put IC cards together.

Then the social-tie, “We will meet and observe social

networks together,” is automatically added to Poly-

phonet Conference.

The kiosk PC sends a user’s action, i.e., viewing

the social network with other users (Action 3), to Ac-

tionLog.

4.4 Tabletop Community

The Tabletop Community is a system based on the

idea of a micro-community similar to that created

around tables and situations where people are drink-

ing coffee in a casual way. Therefore, the environ-

ment that the system is targeting is something similar

to corners of conference rooms and places for coffee

breaks. We have developed the system shown in Fig-

ure 3. An omni-directional camera and RFID card

readers are on the table of the system. Using these

camera and card readers, when the user puts an RFID

card onto a reader, the system detects the individual’s

rough location (direction of the user from the cam-

era) and identiﬁcation of users in the picture, which

is taken by an omni-directional camera, every time a

user places/withdraws a personal RFID card.

When two or three participants put IC cards to-

gether, they can see the network of their photographs

shown in Figure 3.

The Tabletop Community recognizes the user’s ac-

tions when she took photographs and interacted with

other participants (Action 4).

CONTEXT-AWARE WEBLOG TO ENHANCE COMMUNICATION AMONG PARTICIPANTS IN A CONFERENCE

403

Figure 3: Tabletop Community usage.

4.5 Information Clip

Conferences have numerous presentations and

demonstrations, but participants have difﬁculty

recalling details and onsite impressions. We provide

an environment for information clipping by applying

mobile telephones (Figure 4). By scanning a clip

signal, such as a QR code that is printed on the

conference program or RFID tags on a poster, an

application user can record an interesting demon-

stration and can save pictures or comments as an

impression. In addition, users can share their pictures

and comments with presenters.

In the Information Clip, two types of actions are

identiﬁed, i.e., checking a presentation on a confer-

ence program (Action 5) and actually attending a pre-

sentation on site (Action 6).

Figure 4: Information Clip usage.

4.6 Designing Users’ Actions

Through the subsections above, we described the sub-

systems of UbiCoSup. In the ActionLog system,

the participant assistant agent captures users’ actions

from these systems. These six kinds of captured ac-

tions are categorized into two types: actions related

to people and actions related to presentations. Ac-

tions related to people represent communication ac-

tivities. Actions related to presentations encourage

shared opinions about presentations. Table 1 catego-

rizes the captured actions.

5 WEBLOG CONTENT

GENERATION AND

REPRESENTATION

Enhancing communications among participants is a

key to achieving an active community. Action-

oriented experience sharing promotes the exchange

of opinions. In the system, user’s actions are cap-

tured and collected from other information systems.

Based on the context information of the action, the

system generates the Weblog content associated with

the action. The user’s action records are automati-

cally generated through interactions with other sys-

tems and other people in the event space and on the

Web. The user can edit the generated contents. They

are published on the Weblog and are shared among

participants. Exchanging contents among users can

enhance communications. Aggregating the contents

based on their context facilitates context-based access

to the contents; responding to others’ contents engen-

ders topic-based discussion.

5.1 Collecting Users’ Actions,

Generating and Publishing

Weblog Contents

Users’ actions are sent by UbiCoSup subsys-

tems for the ActionLog system via HTTP POST

requests. A subsystem sends a request that

includes queries, such as (UserID, Place,

ActionType, ContextInformation).

ContextInformation depends on its

ActionType, which is pre-registered to Ac-

tionLog. The ActionLog system interprets

the ContextInformation based on its

ActionType, and then it generates the contents.

The ActionLog system generates a Weblog entry

for each action sent from a UbiCoSup subsystem, in-

terpreting the context information of the actions. In-

terpretation of the context information differs accord-

ing to the types of actions. Take a Bookmark ac-

tion for example. A bookmark action is sent from

Polyphonet Conference and its context information is

a presentation ID. The ActionLog agent generates a

Weblog entry using the following steps.

1. Fetch the title and the authors’ names associated to

the Presentation ID.

2. Get a list of users who added the presentation as

a bookmark from Polyphonet Conference (sched-

uler).

3. Get a list of user’s acquaintances from Polyphonet

Conference (social network).

4. Make an intersection of these two lists.

WEBIST 2006 - WEB INTERFACES AND APPLICATIONS

404

Table 1: Types of captured actions.

No. Action System Action Type Context Information

1 Adding a person to the acquaintance list Polyphonet Conference People Participant ID

2 Bookmarking(adding to schedule) a presentation Polyphonet Conference Presentation Presentation ID,

List of Participants who bookmarked the same presentation

3 Viewing the social network among two or three users Kiosk PC People Participant ID

4 tTaking a photograph with other participants Tabletop Community People Participant ID

5 Checking a presentation on a conference program Information Clip Presentation Presentation ID, Userfs Comment

6 Attending a presentation Information Clip Presentation Presentation ID, Userfs Comment

5. Apply the template and generate the content.

An example of the output is shown in Figure 5.

These procedures, the content templates, and the type

deﬁnition of the context information are also pre-

registered to the ActionLog system.

Bookmarked the paper “Paper Title” by Author Name1

(University of X) and Author Name2 (University of Y). Friend Name1,

Friend Name2, and Friend Name3, who are my acquaintances, also

bookmarked this paper.

Figure 5: Example of generated content.

Generated contents are inserted to the user’s We-

blog. The context information, e.g. the presentation

that was attended, the place this action occurred, and

so on, is attached to the contents. A user can edit the

contents on the editor page (Figure 6). She can re-

move or modify the generated sentences, or add new

sentences.

Figure 7 shows a snapshot of the ActionLog system

menu page. On the menu page, the captured actions,

i.e., the generated contents, are listed in the red box.

The user can edit and publish the generated contents.

Figure 8 shows the Weblog-style view, which lists the

user’s published contents in chronological order.

5.2 Aggregating and Sharing

Contents

Efﬁcient sharing of published contents is achieved

through content aggregation. We provide a context-

based aggregation view of contents.

Kaenampornpan et al. (Kaenampornpan, 2004)

proposed the context model for context-aware sys-

tems based on Activity Theory (Engestr

om, 1999).

According to the model, contexts are classiﬁed to six

elements: User, Tools and their availability, Rules,

Community, Division of Labour, Object, and Time.

In explanation of our proposed system, User is the

user herself and the place she is. Tools represent each

UbiCoSup subsystem. Rules are the norms and the so-

cial rules of the conference, and Community includes

the people around the user. Division of Labour is the

role in the action, such as the presenter or attendee.

Object is the target of the action, such as the presenta-

tion that the user attended or the person that the user

added to the acquaintance list. Time is the time at

which the action occurred.

The Weblog-style view is a person-based view of

the contents in chronological order. It includes the

User, Community, and Time items. The Tools, Divi-

sion of Labour and Object items are closely related

to the type of action. We provide aggregation views

on the types and targets of actions. We do not target

Rules in this system.

The basic idea of aggregation is ﬁltering according

to context. The system extracts and lists contents re-

lated to the same presentation (Figure 9). This type of

aggregation treats Object contexts. The system also

aggregates by type of action. In this case, Tools and

Division of Labour are considered. As we previously

mentioned, the personal Weblog-style view includes

User and Community based aggregation. The user can

look through all the contents related to the context that

the user speciﬁes.

5.3 Supporting Communications

Only aggregation is insufﬁcient to enhance commu-

nication among users. Sharing experiences is an im-

portant factor in understanding other people, but af-

ter understanding, we must interact. For this purpose,

we provide the “reply” function. This function allows

users to write new content that includes opinions, im-

pressions, advice, and any other comment about an-

other’s content.

The user can create a replying content by click-

ing the “reply” buttons shown at the bottom of each

section of the content. This function is implemented

based on TrackBack. TrackBack is a typical Weblog

feature that automatically creates reverse links among

contents. The proposed system locates the reply con-

tents using the following reverse links and presenta-

tions.

CONTEXT-AWARE WEBLOG TO ENHANCE COMMUNICATION AMONG PARTICIPANTS IN A CONFERENCE

405

Figure 6: Editor page of ActionLog.

Figure 7: Menu page of ActionLog.

Figure 8: Weblog-style view of ActionLog.

Figure 9: Presentation-based aggregated view of ActionLog.

Furthermore, we provide an aggregated content list

on the editor page (Figure 6). Users can edit their

own actions while reviewing other contents that share

context with the editing action. This capability pro-

motes users’ acquisition of new opinions and infor-

mation about the action or the target of the action.

6 RESULTS

We tested our system at the 7th International Con-

ference on Ubiquitous Computing (UbiComp2005),

which was held on the 12th-14th of September 2005,

in Tokyo. At this conference, there were 122 presen-

tations and 355 authors (including co-authors). About

500 participants joined the conference.

There were 308 Polyphonet users, 175 users regis-

tered from the web system, and 133 users were invited

from information kiosks in the conference room.

Regarding ActionLog, 4586 entries were gener-

ated. There were 3918 entry contents from the ac-

tions related to people and 668 entries for presenta-

tions contents.

Finally, 105 entries were published. Fifteen en-

tries were generated from actions related to people,

64 entries from presentations, and 26 entries were not

generated from actions, i.e., interactions with others,

which included replying entries.

Figure 10 shows the rates of the entries for ac-

tions. Contents related to presentations increased in

rate when published. The rate of replying contents

also occupied a large part. These facts show that a

topic-centered view played an important role in shar-

ing contents among users. Actually, several discus-

sions were made in the ActionLog system.

Figure 10: Rates of entries for actions.

7 RELATED WORK

Several conference support systems have been pro-

posed. IntelliBadge (Cox, 2003) is an RFID-based

academic conference support system that traces users’

positions and provides location search services for

users. It enhances communication at conference sites,

but does not support post-conference communication.

Another system, C-MAP (Sumi, 1998), is a per-

sonal mobile assistance system for exhibition tours. It

WEBIST 2006 - WEB INTERFACES AND APPLICATIONS

406

is intended as a location-aware support by sharing ex-

periences and knowledge using mobile terminals like

PDAs. It provides total support in an event space,

but does not target integrative support of independent

subsystems in the real world or on the Web.

LifeLog research, like MyLifeBits (Gemmell,

2002), is a salient approach to sharing experiences

by recording everything users see and hear. However,

our approach is to share users’ subjective experiences

rather than objective records.

SPECTER (Kr

oner, 2004) is a personal journal

generation system that generates content based on the

users’ action records. This approach closely resem-

bles our approach, but it is speciﬁcally intended to

model users’ behaviors and does not target support-

ing communications or sharing experiences.

8 CONCLUSION

We propose a Weblog with a real-world context called

ActionLog. ActionLog gathers users’ actions, with

related context, and generates entry drafts indeed of

users. With ActionLog, users can accumulate what

they accomplish in small increments as draft entries,

and publish these entries, after they edit generated

drafts. The published entries can be easily aggre-

gated with various aspects because they have context

as metadata.

ActionLog applies to conference support. The Ac-

tionLog Conference worked as an integrator of other

conference support systems. It collects and seam-

lessly integrates users’ actions from the Web and the

real world. The collected actions are shown in various

styles according to their state of use.

The action-oriented integration is easily extensi-

ble. New subsystems are pluggable through “action.”

This means that our approach can be applied in vari-

ous situations. Future work includes designing a sys-

tem with more proactive supports. The tested imple-

mentation is a limited service in terms of time and

place. When we utilize the system in an open envi-

ronment, the information that will become available

will be enormous so that a user cannot manage them.

ACKNOWLEDGEMENTS

The Polyphonet Conference has been supported by

the New Energy and Industrial Technology Develop-

ment Organization (NEDO) as the 04A11502a project

ID. The cellular phones with RFID tag readers for the

conference were courtesy of KDDI Corporation. The

IC card reader system and technology are supported

by Synergymedia Corporation.

REFERENCES

Avesani, P., Cova, M., Hayes, C., and Massa, P. (2005).

Learning Contextualised Weblog Topics. In Proc. of

WWW2005 2nd WS on the Weblogging Ecosystem:

Aggregation, Analysis and Dynamics.

Cox, D., Kindratenko, V., and Pointer, D. (2003). Intel-

liBadge: Towards Providing Location-Aware Value-

Added Services at Academic Conferences. In Proc.

of 5th Int’l Conf. on Ubiquitous Computing (Ubi-

Comp2003).

Duckham, M. and Kulik, L. (2005). A formal Model of

Obfuscation and Negotiation for Location Privacy. In

Proc. of 3rd Int’l Conf. on Pervasive Computing (Per-

vasive2005).

Engestr

om, Y. (1999). Activity theory and individual and

social transformation. In Y. Engestrom, R. Miettinen

and R. Punamaki (eds.), Perspectives on Activity The-

ory, Cambridge University Press.

Gemmell, J., Bell, G., Lueder, R., Drucker, S., and Wong,

C. (2002). MyLifeBits: fulﬁlling the Memex vision.

In Proc. of 10th ACM Int’l Conf. on Multimedia.

Gruteser, M. and Grunwald, D. (2003). Anonymous usage

of location-based services through spatial and tempo-

ral cloaking. In Proc. of 1st Int’l Conf. on Mobile Sys-

tems, Applications and Services (MobiSys’03).

Kaenampornpan, M. and O’Neill, E. (2004). Modelling

Context: An Activity Theory Approach. In Proc.

of 2nd European Symposium on Ambient Intelligence

(EUSAI2004).

Karger, D. and Quan, D. (2004). What Would It Mean to

Blog on the Semantic Web? In Proc. of 3rd Int’l Se-

mantic Web Conf. (ISWC2004).

oner, A., Baldes, S., Jameson, A., and Bauer, M. (2004).

Using an Extended Episodic Memory Within a Mo-

bile Companion. In Proc. of Pervasive 2004 WS on

Memory and Sharing of Experience (MSE2004).

Matsuo, Y., Tomobe, H., Hasida, K., and Ishizuka, M.

(2003). Mining Social Network of Conference Par-

ticipants from the Web. In Proc. of 2003 IEEE/WIC

Int’l Conf. on Web Intelligence (WI2003).

Newman, W., Eldridge, M., and Lamming, M. (1991).

PEPYS: Generating Autobiographies by Auto-

matic Tracking. In Proc. of 2nd European Conf.

on Computer-Supported Cooperative-Work (EC-

SCW’91).

Sumi, Y., Etani, T., Fels, S., Simonet, N., Kobayashi, K.,

and Mase, K. (1998). C-MAP: Building a Context-

Aware Mobile Assistant for Exhibition Tours. In Toru

Ishida (Ed.), Community Computing and Support Sys-

tems: Social Interaction in Networked Communities,

LNCS 1519, Springer.

Sumi, Y., Sakamoto, R., Nakao, K., and Mase, K. (2002).

ComicDiary: Representing Individual Experiences in

a Comics Style. In Proc. of 4th Int’l Conf. on Ubiqui-

tous Computing (UbiComp2002).

CONTEXT-AWARE WEBLOG TO ENHANCE COMMUNICATION AMONG PARTICIPANTS IN A CONFERENCE

407