Collaboration Technologies and Social

Computing 10th International

Conference CollabTech 2018 Costa de
Caparica Portugal September 5 7 2018
Proceedings Hironori Egi
Visit to download the full and correct content document:
https://textbookfull.com/product/collaboration-technologies-and-social-computing-10th
-international-conference-collabtech-2018-costa-de-caparica-portugal-september-5-7-
2018-proceedings-hironori-egi/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Computational Collective Intelligence 10th

International Conference ICCCI 2018 Bristol UK
September 5 7 2018 Proceedings Part I Ngoc Thanh Nguyen

https://textbookfull.com/product/computational-collective-
intelligence-10th-international-conference-iccci-2018-bristol-uk-
september-5-7-2018-proceedings-part-i-ngoc-thanh-nguyen/

Computational Collective Intelligence 10th

International Conference ICCCI 2018 Bristol UK
September 5 7 2018 Proceedings Part II Ngoc Thanh
Nguyen
https://textbookfull.com/product/computational-collective-
intelligence-10th-international-conference-iccci-2018-bristol-uk-
september-5-7-2018-proceedings-part-ii-ngoc-thanh-nguyen/

Technological Innovation for Industry and Service

Systems 10th IFIP WG 5 5 SOCOLNET Advanced Doctoral
Conference on Computing Electrical and Industrial
Systems DoCEIS 2019 Costa de Caparica Portugal May 8 10
2019 Proceedings Luis M. Camarinha-Matos
https://textbookfull.com/product/technological-innovation-for-
industry-and-service-systems-10th-ifip-wg-5-5-socolnet-advanced-
doctoral-conference-on-computing-electrical-and-industrial-
systems-doceis-2019-costa-de-caparica-portugal-m/

Reversible Computation 10th International Conference RC

2018 Leicester UK September 12 14 2018 Proceedings
Jarkko Kari

https://textbookfull.com/product/reversible-computation-10th-
international-conference-rc-2018-leicester-uk-
september-12-14-2018-proceedings-jarkko-kari/
Emerging Technologies for Authorization and
Authentication First International Workshop ETAA 2018
Barcelona Spain September 7 2018 Proceedings Andrea
Saracino
https://textbookfull.com/product/emerging-technologies-for-
authorization-and-authentication-first-international-workshop-
etaa-2018-barcelona-spain-september-7-2018-proceedings-andrea-
saracino/

Broadband Communications, Networks, and Systems: 9th

International EAI Conference, Broadnets 2018, Faro,
Portugal, September 19–20, 2018, Proceedings Victor
Sucasas
https://textbookfull.com/product/broadband-communications-
networks-and-systems-9th-international-eai-conference-
broadnets-2018-faro-portugal-september-19-20-2018-proceedings-
victor-sucasas/

Quantitative Evaluation of Systems 15th International

Conference QEST 2018 Beijing China September 4 7 2018
Proceedings Annabelle Mciver

https://textbookfull.com/product/quantitative-evaluation-of-
systems-15th-international-conference-qest-2018-beijing-china-
september-4-7-2018-proceedings-annabelle-mciver/

Frontiers in Cyber Security First International

Conference FCS 2018 Chengdu China November 5 7 2018
Proceedings Fagen Li

https://textbookfull.com/product/frontiers-in-cyber-security-
first-international-conference-fcs-2018-chengdu-china-
november-5-7-2018-proceedings-fagen-li/

Intelligent Human Computer Interaction 10th

International Conference IHCI 2018 Allahabad India
December 7 9 2018 Proceedings Uma Shanker Tiwary

https://textbookfull.com/product/intelligent-human-computer-
interaction-10th-international-conference-ihci-2018-allahabad-
india-december-7-9-2018-proceedings-uma-shanker-tiwary/
 Hironori Egi · Takaya Yuizono
Nelson Baloian · Takashi Yoshino
Satoshi Ichimura · Armanda Rodrigues (Eds.)
LNCS 11000

Collaboration Technologies
and Social Computing
10th International Conference, CollabTech 2018
Costa de Caparica, Portugal, September 5–7, 2018
Proceedings

123
Lecture Notes in Computer Science 11000
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zurich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology Madras, Chennai, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7407
Hironori Egi Takaya Yuizono
•

Nelson Baloian Takashi Yoshino

•

Satoshi Ichimura Armanda Rodrigues (Eds.)

•

Collaboration Technologies
and Social Computing
10th International Conference, CollabTech 2018
Costa de Caparica, Portugal, September 5–7, 2018
Proceedings

123
Editors
Hironori Egi Takashi Yoshino
Department of Informatics Wakayama University
The University of Electro-Communications Wakayama City
Chofu, Tokyo Japan
Japan
Satoshi Ichimura
Takaya Yuizono Otsuma Women’s University
Graduate School of Knowledge Science Tokyo
Japan Advanced Institute of Science Japan
and Technology
Nomi, Ishikawa Armanda Rodrigues
Japan Departamento de Informatica
Universidade Nova de Lisboa
Nelson Baloian Caparica
University of Chile Portugal
Santiago
Chile

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-319-98742-2 ISBN 978-3-319-98743-9 (eBook)
https://doi.org/10.1007/978-3-319-98743-9

Library of Congress Control Number: 2018950648

LNCS Sublibrary: SL1 – Theoretical Computer Science and General Issues

© Springer Nature Switzerland AG 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

Message from the General Chairs

We are delighted with the brilliant success of the 10th International Conference on
Collaboration Technologies (CollabTech 2018). CollabTech 2018 offered a unique
forum for academics and practitioners to present and discuss innovative ideas, methods,
or implementations related to collaboration technologies, whose contributions to the
successful completion of various routine collaboration activities have been enhanced
by recent advances in networking, computing, and interaction technologies.
CollabTech conferences were held in Tokyo in 2005, Tsukuba in 2006, Seoul in
2007, Wakayama in 2008, Sydney in 2009, Sapporo in 2012, Santiago in 2014,
Kanazawa in 2016, and Saskatoon in 2017. Following the success of the joint orga-
nization of CollabTech conferences with CRIWG 2014, CRIWG 2016, and CRIWG
2017, CollabTech 2018 was co-located and organized with CRIWG 2018 in Costa de
Caparica, Portugal. We believe that our selection of this venue guaranteed the success
of this technical conference, which was enriched by the culture and scenery of Portugal.
Although the CRIWG and CollabTech communities had similar research topics and
goals, they had been geographically located in different regions. Therefore, we believed
this joint endeavor would provide an interesting opportunity for the two communities
to meet and get to know each other.
As the conference chairs of CollabTech 2018, we know that the success of the
conference ultimately depends on the efforts of many people who worked with us in
planning and organizing the conference. We thank the program co-chairs for their wise
counsel and brilliant suggestions regarding the organization of the Program Committee
to ensure that it conducted a thorough and timely review of papers, and our sponsors,
who helped us to make CollabTech 2018 affordable for all its participants. In addition,
we attribute the success of the conference to the efforts of Universidade NOVA de
Lisboa, the Special Interest Group (SIG) on Groupware and Network Services of the
IPSJ, the SIG on Cyberspace of the Virtual Reality Society of Japan, and the SIG on
Communication Enhancement of the Human Interface Society.
Our technical program was diverse and encompassed approximately 16 technical
papers. Further, we provided the participants with numerous opportunities for informal
networking. We are pleased that the conference was fruitful for all the participants and
significantly contributed to the development of academic interest in this research field.

September 2018 Takashi Yoshino

Satoshi Ichimura
Armanda Rodrigues
 Message from the Program Chairs

After nine events of the International Conference on Collaboration Technologies series,

we had the tenth edition (CollabTech 2018) in Costa de Caparica, Portugal. The
following topics on collaboration technologies were discussed:
– Communication Enhancement
– Inter-Cultural Collaboration
– Learning Support System
– Entertainment System
– Social Studies
– UI and UX
For this conference, we received 36 submissions (26 full papers, ten work-in-progress
papers) and assigned three reviewers per full paper or two reviewers per work-in-progress
paper. As a result, we had 12 full papers, and four work-in-progress papers. The
acceptance rate was 44%. Because of the high quality of submissions, many excellent
papers were not among those accepted. We hope that the detailed technical review
comments we provided are helpful.
Without our distinguished Program Committee members, we could not have
maintained our high standards. We truly appreciated their devotion. Finally, we hope
that these proceedings serve as a reference for future researchers in this rapidly
evolving field.

September 2018 Hironori Egi

Takaya Yuizono
Nelson Baloian
 Organization

Conference Co-chairs
Takashi Yoshino Wakayama University, Japan
Satoshi Ichimura Otsuma Women’s University, Japan
Armanda Rodrigues Universidade NOVA de Lisboa, Portugal

Program Co-chairs
Hironori Egi University of Electro-Communications, Japan
Takaya Yuizono Japan Advanced Institute of Science and Technology,
Japan
Nelson Baloian Universidad de Chile, Chile

Publication Chair
Junko Ichino Tokyo City University, Japan

IPSJ SIG GN Liaison

Noriaki Saito Tokyo Online University, Japan

VRSJ SIG CS Liaison

Kazuyuki Iso NTT, Japan

HIS SIG CE Liaison

Takashi Yoshino Wakayama University, Japan

Steering Committee
Hideaki Kuzuoka University of Tsukuba, Japan
Ken-ichi Okada Keio University, Japan
Jun Munemori Wakayama University, Japan
Minoru Kobayashi Meiji University, Japan
Hiroaki Ogata Kyoto University, Japan
Tomoo Inoue University of Tsukuba, Japan
X Organization

Program Committee
Gwo-Dong Chen National Central University, Taiwan
Hui-Chun Chu Soochow University, Taiwan
Kinya Fujita Tokyo University of Agriculture and Technology,
Japan
Atsuo Hazeyama Tokyo Gakugei University, Japan
Gwo-Jen Hwang National Taiwan University of Science
and Technology, Taiwan
Tomoo Inoue University of Tsukuba, Japan
Yutaka Ishii Okayama Prefectural University, Japan
Kazuyuki Iso NTT, Japan
Marc Jansen University of Applied Sciences Ruhr West, Germany
Jongwon Kim Gwangju Institute of Science and Technology,
South Korea
Hyungseok Kim Konkuk University, South Korea
Wim Lamotte Hasselt University, Belgium
Yuan Tian Singapore Management University, Singapore
Chen-Chung Liu National Central University, Taiwan
Wolfram Luther University of Duisburg-Essen, Germany
Hideyuki Nakanishi Osaka University, Japan
Mamoun Nawahdah Birzeit University, Palestine
Masayuki Okamoto Toyota, Japan
Masaki Omata University of Yamanashi, Japan
Nobuchika Sakata NARA Institute of Science and Technology, Japan
Yoshiaki Seki Tokyo City University, Japan
Hidekazu Shiozawa Tamagawa University, Japan
Daniel Spikol Malmo University, Sweden
Hao-Chuan Wang National Tsing Hua University, Taiwan
Kazushi Nishimoto Japan Advanced Insutitute of Science and Technology,
Japan
Shin Takahashi University of Tsukuba, Japan
Satoshi Nakamura Meiji University, Japan
 Contents

Communication Enhancement

Discussion Map with an Assistant Function for Decision-Making:

A Tool for Supporting Consensus-Building . . . . . . . . . . . . . . . . . . . . . . . . 3
Ryunosuke Kirikihira and Kazutaka Shimada

An Integrated Support System for Disaster Prevention Map-Making

Using Town-Walk Information Gathering. . . . . . . . . . . . . . . . . . . . . . . . . . 19
Sojo Enokida, Takashi Yoshino, Taku Fukushima, Kenji Sugimoto,
and Nobuyuki Egusa

Concealment-Type Disaster Prevention Information System Based

on Benefit of Inconvenience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Satoko Shigaki and Takashi Yoshino

Development of a Stroll Support System Using Route Display

on a Map and Photograph Sharing Service . . . . . . . . . . . . . . . . . . . . . . . . . 48
Junko Itou, Takaya Mori, Jun Munemori,
and Noboru Babaguchi

Inter-Cultural Collaboration

Machine Translation Usage in a Children’s Workshop . . . . . . . . . . . . . . . . . 59

Mondheera Pituxcoosuvarn, Toru Ishida, Naomi Yamashita,
Toshiyuki Takasaki, and Yumiko Mori

Learning Support System

A Presentation Supporting System for Programing Workshops

for Elementary School Students . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Koki Ito, Maki Ichimura, and Hideyuki Takada

Gamifying the Teaching and Learning Process in an Advanced

Computer Programming Course . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Mamoun I. Nawahdah

Designing a System of Generating Sound Environment for Promoting

Verbal Communication in Classroom. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Riri Sekine, Yasutaka Asai, and Hironori Egi
XII Contents

“Discuss and Behave Collaboratively!” – Full-Body Interactive Learning

Support System Within a Museum to Elicit Collaboration with Children . . . . 104
Mikihiro Tokuoka, Hiroshi Mizoguchi, Ryohei Egusa,
Shigenori Inagaki, Fusako Kusunoki, and Masanori Sugimoto

Entertainment System

Can Social Comments Contribute to Estimate Impression

of Music Video Clips?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Shunki Tsuchiya, Naoki Ono, Satoshi Nakamura,
and Takehiro Yamamoto

Detection of Football Spoilers on Twitter . . . . . . . . . . . . . . . . . . . . . . . . . . 129

Yuji Shiratori, Yoshiki Maki, Satoshi Nakamura,
and Takanori Komatsu

Social Studies

Analysis of Facilitators’ Behaviors in Multi-party Conversations

for Constructing a Digital Facilitator System . . . . . . . . . . . . . . . . . . . . . . . 145
Tsukasa Shiota, Takashi Yamamura, and Kazutaka Shimada

Consideration of a Method to Support Face-to-Face Communication

Using Printed Stickers Featuring a Picture of a Character
Expressing a Mood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Yuri Nishimura and Minoru Kobayashi

UI and UX

Migaco: Supporting Young Children’s Tooth Brushing

with Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Satoshi Ichimura

Improving Visibility and Reducing Resistance of Writers

to Fusion of Handwritten and Type Characters . . . . . . . . . . . . . . . . . . . . . . 185
Mikako Sasaki, Junki Saito, and Satoshi Nakamura

PopObject: A Robotic Screen for Embodying Video-Mediated

Object Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Kana Kushida and Hideyuki Nakanishi

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

Communication Enhancement
 Discussion Map with an Assistant
Function for Decision-Making: A Tool
for Supporting Consensus-Building

Ryunosuke Kirikihira and Kazutaka Shimada(B)

Department of Artificial Intelligence, Kyushu Institute of Technology,

680-4 Kawazu Iizuka, Fukuoka 820-8502, Japan
{r kirikihira,shimada}@pluto.ai.kyutech.ac.jp

Abstract. In this paper, we propose a tool for supporting consensus-

building in conversations with multiple participants. We call it “Dis-
cussion Map with Assistant (DMA)”. It consists of nodes and links.
We classify the nodes into two types; alternatives and criteria. Alter-
natives represent what the participants are choosing between. Criteria
are used to judge the alternatives. Each criterion contains an importance
value. Each link between nodes also contains an importance value. The
system estimates a ranking list of alternatives among participants from
each map. We introduce a forgetting function to the model. The system
also supports the decision-making process by using discussion maps from
participants. It generates sentences and charts that describe the current
state of the discussion. We evaluate the effectiveness of the discussion
map system with DMA in a decision-making task experimentally.

Keywords: Discussion map · Decision support system

Consensus estimation · Support with charts and sentences

1 Introduction

Supporting consensus-building in conversations with multiple participants is a

very important task in intelligent systems. That is expected to be applied in a
wide range of fields. Participants in discussion often struggle to identify the most
suitable solution for a decision on a meeting agenda because there are generally
many alternatives and criteria related to making the decision. It is important
for the participants to visualize the discussion state for making a good decision.
The visualization also helps each participant to understand temporal opinions.
Such a supporting system for consensus-building can play a very impor-
tant role in many situations. For example, in education, problem-based learning
(PBL) has recently become a highly regarded approach to learning [6]. Discus-
sion among participants has a critical role in a PBL environment. However,
students tend to be limited, not very effective and often failing to make a sat-
isfying decision. It leads to the failure of the discussion. To conduct smooth,
c Springer Nature Switzerland AG 2018
H. Egi et al. (Eds.): CollabTech 2018, LNCS 11000, pp. 3–18, 2018.
https://doi.org/10.1007/978-3-319-98743-9_1
4 R. Kirikihira and K. Shimada

Overhead-view camera

Discussion Map

Microphone and
Multi-camera images

Roles of Discussion Map

1. Management of personal
opinions and information
sharing
2. Feedback for consensus-
building

Fig. 1. Discussion map for decision-making tasks.

active and productive discussions, they need a facilitator who controls the dis-
cussion appropriately. However, it is impractical to assign a good facilitator to
each group in the PBL environment because of lack of human resources. Similar
cases also appear in other situations, such as business meetings and group dis-
cussions. Although a project manager needs to appropriately handle a discussion
in business meetings, he/she might not have a remarkable skill in terms of dis-
cussion facilitation. Ordinary people in a group discussion might subconsciously
need help from others to generate a good decision. Therefore, supporting such
discussions is an important task.
In this paper, we propose a tool for supporting consensus-building in conver-
sations with multiple participants in various environments, such as meetings and
PBL situations. The goal of our study is to construct a system that supports
consensus-building and management of conversation for high-quality discussions.
We call it “digital facilitator”. Figure 1 shows the outline of our system. Here
we focus on laptop PCs or tablet terminals as an input tool for discussions. The
prototype system in this paper is developed as a Web application. We propose
a discussion map which consists of nodes and links among nodes constructed by
each participant. Participants can express their ideas and opinions by using the
discussion map. Each discussion map is a kind of visualization of each partici-
pant’s thought. The visualization helps to easily understand own thoughts. In
addition, discussion maps are manageable for computer systems because they are
based on a graph structure, as compared with natural language texts, images,
and sounds. Our system estimates the discussion state and a latent consensus
from discussion maps of each participant, and then explains the circumstances
to each participant. This paper contains two contributions;
(1) Our system estimates individual opinions of participants by using a scoring
method with a decay factor.
(2) Our method explains the discussion state to participants by using charts
and sentences, namely a decision-making assistant function (DMA).
We evaluate these two points through experiments in this paper.
 Discussion Map with an Assistant Function for Decision-Making 5

2 Related Work
There are many fields related to our method. One of the targets in our research
is the PBL environment. Several researchers have proposed collaborative learn-
ing support systems and approaches based on constructive interaction [11,15].
Utilizing graph structures, such as a concept map, is one of the most effective
approaches for education supports. Villalon and Calvo [19] have reported a defi-
nition and a framework for its evaluation of concept map mining. Yamasaki et al.
[20] proposed the Kit-Build method based on a concept map. The purposes of
these studies were to extract a concept map automatically or to compare the
goal map and each learner map. In contrast, our purpose is to manage opinions
and estimate the consensus of participants in a discussion. Takagi and Shimada
[18] have proposed a system for collaborative learning with tablet terminals.
However, their purpose was just to estimate a level of understanding among
participants. Suzuki et al. [17] have proposed a collaborative learning tool using
tablet terminals, XingBoard. Masukawa [10] has proposed a web-based note-
book for collaborative learning. The purposes of these studies were to construct
a framework to support the discussion environment. On the other hand, the goal
of our study is to develop a digital facilitator based on a computer and support
interaction itself.
The target of our research is not limited to the PBL situation. El-Assady et al.
[4] have developed a visual analytics framework for multi-party discourse data.
However, the purpose of this study was to visualize transcribed text interactively
for content analysis. Ito et al. [7] have developed an open on-line workshop system
called COLLAGREE that has facilitator support functions for an internet-based
town meeting. They incorporated an incentive mechanism for large-scale collec-
tive discussions to the system [8]. The system is effective. However, the purpose
of this system is to gather opinions from many participants on the web and
extract discussion points from them. The purpose of our system is to visualize
real-time thoughts and opinions of participants by using discussion maps, and
then support the decision-making. Katsura et al. [9] have proposed an argu-
mentation support tool which displays justified arguments to participants. The
purpose was to judge whether an argument is logically correct or not. Our target
does not always require the correctness of the consensus in a discussion. Nagao
et al. [13] have proposed a system for discussion mining. The system generated
meeting summaries semi-automatically, retrieved the discussion contents, and
generated an answer to a certain question based on the accumulated discussion
contents. Nagao [12] also has proposed a creative activity support system. These
systems are useful and the motivation is similar to our purpose. One major issue
of these systems was to support task execution and evaluation results, namely
the plan-do-check-act (PCDA) cycle. On the other hand, we focus on support-
ing a discussion by an assistant function based on our discussion map system for
making the best decision.
VatueCharts by [2] and LineUp by [5] are good tools to make the best decision
from several alternatives and criteria. Participants can add weight to the criteria,
and then easily identify the best choice by using these tools. Alonso et al. [1]
6 R. Kirikihira and K. Shimada

Fig. 2. The outline of our discussion map system with assistant functions. Each map
shares nodes. However, links are different. Therefore, each participant doesn’t know
other participants’ maps. Only the system knows the whole situation and generates
sentences and charts for decision-making success.

have proposed a consensus visualization method based on a clustering algorithm

with several measures. It generated a consensus diagram based on opinions of
each participant. The motivation of these studies is to visualize the situation
and consensus. On the other hand, the main purpose of our study is to support
discussions by providing advice that is estimated from the current discussion
maps.

3 Discussion Map System with Assistant

In this section, we explain our discussion map system. First, we describe the dis-
cussion map (DM) itself. Then, we explain a consensus estimation function from
DMs. We introduce a decay factor to the function. Finally, we describe a facil-
itator function, namely Discussion Map with Assistant (DMA). DMA explains
the current discussion state to participants and provides advice that is esti-
mated from the current DMs of all participants. Figure 2 shows the outline of
the system.
 Discussion Map with an Assistant Function for Decision-Making 7

Add a node Estimation Add a node Estimation

The DM of participant P1 The DM of participant P2

Fig. 3. Two DMs from two participants (P1) and (P2). Each participant expresses own
thoughts on own map. Other participants cannot see the map. (Color figure online)

The system pops up the menu when a participant selects a criterion.

The left one is a link button to an alternative. There are 3 grades; low, middle and high.
The middle one is a button to the weighting.
The right one is a button to delete the node.

Fig. 4. The interface of the weighting to a criterion.

3.1 Discussion Map

The purpose of DMs is to visualize concepts, ideas, and opinions of each partic-
ipant. Each DM consists of several nodes and links. Figure 3 shows an example
of DMs from two participants, P1 and P2, about “Which programming envi-
ronment is the best for Web applications?” Nodes are classified into two roles;
alternatives and criteria. Alternatives are displayed as blue nodes and represent
what the participants are choosing between. In the figure, “slim” and “rails” are
alternatives created by participants in the discussion. Criteria are displayed as
green nodes and are used to judge the alternatives. In the figure, “easy-to-use”
and “# of documents” are criteria. Each criterion has a weight as an importance
value. The importance value has three grades; high, middle and low. They are
expressed by gradations of color on the map; deep green for “high” and light
green for “low.” Each participant assigns its own importance value to each cri-
terion. In the figure, the participant who created the DM thinks that “# of
documents” is more important than “easy-to-use.” Figure 4 shows the interface
to weight a criterion.
8 R. Kirikihira and K. Shimada

Value

Comments

Fig. 5. The interface of the weighting to a link.

An alternative often has some links to criteria1 . A link specifies a 5-grade

evaluation value between an alternative and a criterion; −−, −, 0, + and ++.
Each participant expresses the importance of each link by using the evaluation
value. In the DM (P1) of Fig. 3, the participant thinks that “slim” is very good
in terms of “easy-to-use” while “rails” is weak. Figure 5 shows the interface to
set the weight to a link.
We now explain the concept of DMs in details. In DMs, nodes, namely alter-
natives and criteria, are shared among participants. On the other hand, links
depend on each participant. Therefore, the structures of DMs created by each
participant are different. Please see Fig. 3 again. It contains two DMs from two
participants P1 and P2. From the DMs, we can infer the following points;
– P1 attaches importance to “# of documents”
– P2 attaches importance to “easy-to-use”
– P2 probably does not attach importance to “# of documents” because the
weight is low and there is no link to any alternatives.
– The evaluation values for “easy-to-use” for “slim” exhibit a similar tendency
for both P1 and P2; ++ and +, namely positive evaluation.
From these points, our system can estimate that the consensus (the best choice)
should be “slim” in this discussion.

3.2 Consensus Estimation

The purpose of the DM system is to support consensus-building by using each
map of participants. For the purpose, we need to estimate participants’ opinions
and thoughts from DMs. In other words, which alternative is preferred among
participant? If the system can estimate some alternatives with high importance
in the discussion, it can support to make the decision in the last stage of the
discussion by using the estimated results. For example, the system can say “the
current ranking of alternatives is as follows; 1st: A3 , 2nd: A1 and 3rd: A2 .” The
information is useful for the participants to make a decision. In this section, we
explain a rank estimation model of alternatives as a consensus estimation model.
1
Note that not all alternatives have a link. It depends on the participant that creates
the DM.
 Discussion Map with an Assistant Function for Decision-Making 9

The basic idea is based on the summation of the evaluation value between
an alternative and each criterion. The score of an alternative Aj of a participant
Pi is computed as follows:
N

Score(Pi , Aj ) = wCk × ev(Pi , Aj , Ck ) (1)
k

where Ck is a criterion on the map. N is the number of criteria on the map. wCk
is the weight of Ck , namely the importance value of a node. We set 1.0, 0.5 and
0.1 to “High”, “Middle” and “Low” in Fig. 4. ev(Pi , Aj , Ck ) is the evaluation
value between Aj and Ck of Pi , namely the value of a link between nodes (5
grade scores; 0, 2, 4, 8 and 10 for −−, −, 0, + and ++ in Fig. 5). These values
are determined heuristically.
Here we consider a temporal characteristic of discussion. Participants need to
generate alternatives and criteria on the basis of divergent thinking in the early
stage of the discussion. Then, alternatives with low importance are culled while
alternatives with high importance are discussed continually. In other words, in
the current stage of a discussion, a topic that is discussed continually is more
important than a topic that is discussed sporadically. In our DM system, we
assume that Aj and Ck that each participant operates frequently are important.
To incorporate this characteristic to the score, we introduce a forgetting function
[3]. Ebbinghaus has formulated a current memory level b after t minutes2 as
follows:
100k
b= (2)
(log10 t)c + k
where c and k are constant; c = 1.25 and k = 1.84 in [3]. From the Eq. (2), we
compute the saving level sv on a current t as follows:
b
sv = (3)
100
We apply the saving level sv into the Eq. (1).
N

M emScore(Pi , Aj ) = wCk × ev(Pi , Aj , Ck ) × sv(Aj ) × sv(Ck ) (4)
k

where sv(Aj ) and sv(Ck ) denote the saving levels of Aj and Ck . The t of these
saving levels is the elapsed time since the last participant’s action to Aj or Ck .
Here the action denotes

– the creation of Aj or Ck ,
– the change of the evaluation value of a link between Aj and Ck ,
– the change of the importance value of Ck , or
– the move of Aj or Ck on the discussion map.

2
The initial memory level is 100 in this formulation.
10 R. Kirikihira and K. Shimada

Alternatives (DMA)

Display overall and user’s evaluation

Overall You
Current state

AKIYOSHI-DAI
YUHUIN AKIYOSHI-DAI and YUHUIN obtain high scores.
AKIYOSHI-DAI and YUHUIN are valued in terms of “summer-like”
and “price”, respectively.
Feedback on overall opinion

Which is better in terms of “distance”, AKIYOSHI-DAI or YUHUIN?

How about discussing “summer-like”?

Alternatives with criteria Feedback on user’s opinion

Radar chart on criteria

Radar chart on alternatives You value

Fig. 6. The overview of DMA (DM assistant).

The final score of an alternative Aj is computed as follows:

Np
i M emScore(Pi , Aj )
F S(Aj ) = (5)
Np

where Np is the number of participants in the discussion.

3.3 Discussion Map with Assistant: DMA

Our system contains a support function for discussions, Discussion Map with
Assistant (DMA). The purpose of DMA is to support divergent thinking of each
participant in the middle stage of a discussion and consensus-building in the last
stage of the discussion. DMA generates some charts and sentences as a current
summary in a discussion. Charts and sentences are generated from discussion
maps of each participant on the basis of the consensus score computed by Eq. (5)
in Sect. 3.2.
Figure 6 shows the interface of DMA. This example is an output of a par-
ticipant about a discussion topic, “Where is the best location for a travel with
laboratory members?” The left side of DMA consists of four types of charts3 .
They are opened and closed by using the icons “+” and “−”, respectively. The
right side of DMA consists of three sentence parts; a summary of the consensus
candidates on the current state, feedback based on alternatives and criteria with
high scores from all participants and feedback based on alternatives and criteria
with high scores from the user that sees this DMA interface.
The roles of each chart and sentence are as follows:

3
Note that three of them are hidden in this figure.
 Discussion Map with an Assistant Function for Decision-Making 11

Fig. 7. Charts generated from DMA.

1. Overall and user’s evaluation (the top-left in Fig. 6):

The purpose of this chart is to understand the whole discussion and own
opinions. In addition, the user understands the gap between the majority
view and his/her opinion. For example, in Fig. 6, the user’s preference, namely
HUIS TEN BOSCH, is not valued by other participants while a candidate
AKIYOSHI-DAI is valued by other participants.
2. Alternatives with criteria ((a) in Fig. 7):
This chart is used in the last stage of the discussion. Some alternatives often
contain high scores in the last stage. As a result, participants struggle to
determine the final decision. In this situation, each user can understand the
evaluation points of each alternative by using this chart. For example, assume
that the participants need to select AKIYOSHI-DAI or YUHUIN in (a) of
Fig. 7. From the chart, we can understand that the criterion “distance” is
the approximately same value. In other words, this situation implies that
“distance” may be the most important criterion to select one from two alter-
natives because it becomes clear about comparative merits and demerits of
two alternatives if the criterion has a difference between them. This becomes
a good trigger for the discussion. In addition, DMA displays his/her ratio of
criteria of each alternative by clicking the button on the topside, “overall”
and “you”. By using this own chart, he/she can organize his/her thoughts.
3. Radar chart on criteria ((b) in Fig. 7):
Participants can easily compare all criteria about selected alternatives
by using this chart. In this example, the most different points between
AKIYOSHI-DAI and YUHUIN are visualized, namely “summer-like” vs.
“price” in two alternatives.
12 R. Kirikihira and K. Shimada

4. Radar chart on alternatives ((c) in Fig. 7):

Participants often want to know evaluation values of all alternatives when
they discuss some criteria. This chart is used for the situation. Each par-
ticipant can easily understand the difference of all alternatives on selected
criteria. In this example, the score of TSUNOSHIMA about “distance” is
higher than the others and that of YUHUIN about “price” is higher than the
others.
5. Sentence feedback to the user (the right side in Fig. 6):
This process is based on a template-based sentence generation scheme.
For the explanation of the current state, our system selects the top alternative
based on the score, Eq. (5), first. Then, it searches another alternative that is
close4 in the score to the best alternative. The templates are:
– [A1 ] and [A2 ] obtain high scores.
– [A1 ] and [A2 ] are valued in terms of [C1 ] and [C2 ], respectively.
where Ai is an alternative and Ci is a criterion related to the Ai .
The purpose of the feedback on the overall opinion is to pull the trigger that
leads to a good final decision. Therefore, our system suggests a topic, namely
a good criterion, for two alternatives. The template is:
– Which is better in terms of [Cvar ], [A1 ] or [A2 ]?
where Cvar is a criterion that links to both of the A1 and A2 and the variance
of the score is the smallest in all criteria. To engage in wide-ranging discussion,
our system also suggests another criterion that is barely mentioned in the
discussion if such a criterion exists in the discussion. The template is:
– How about discussing [Clow ]?
where Clow is the number of links to alternatives that is the lowest in the dis-
cussion maps. In this regard, however, the criterion must link to more than
half of the alternatives. If any conditions mentioned above for feedback on
the overall opinion are not satisfied, the system outputs the sentence “Let’s
keep up the discussion.”
Our system also generates feedback to individuals. It suggests the gap
between his/her preference and the current consensus by using the follow-
ing templates:
– You value [Aown ] highly although the current consensus is [A1 ].
– How about providing comments about [Cgap1 ] and [Cgap2 ] to other par-
ticipants?
where Aown is the best alternative of the user (Aown = A1 ). Cgap1 and Cgap2
are criteria with high scores in the user and the scores are smaller than scores
that computed from all participants. Our system also generates “Let’s keep
up the discussion.” if these conditions are not satisfied.

4 Experiment
In this paper, we evaluate two parts; the accuracy of consensus estimation
explained in Sect. 3.2 and the effectiveness of DMA explained in Sect. 3.3.
4
The difference of the score is 5% or less.
 Discussion Map with an Assistant Function for Decision-Making 13

Table 1. The final numbers of alternatives and criteria.

Method D1 D2 D3 D4
Alternatives 7 9 12 6
Criteria 5 5 5 4

Table 2. The experimental result: RMSE for all ranks.

Method D1 D2 D3 D4 Ave
Baseline 0.76 2.26 1.58 0.82 1.35
Proposed 0.00 2.11 1.58 0.58 1.07

4.1 Evaluation of Consensus Estimation

We evaluated the consensus estimation method with four real discussions (D1,
D2, D3, and D4). Each discussion consisted of four participants for 20 min. The
topics of each discussion were
(T1) Where is the best location for a seminar camp?: D1 and D2
(T2) What store/restaurant/cafe do you want in this university?: D3 and D4
Each participant created a DM in the discussion. The statistics of DMs in each
discussion are shown in Table 1. The participants made the decision at the end
of the discussion. After the discussion, each participant ranked each alternative.
We evaluated our method with the root mean square error value between the
system ranks by the consensus estimation method and participants’ ranks, as
follows: 

1  N
RM SE =  (SysRanki − P artRanki )2 (6)
N i=1

where N is the number of alternatives in the discussion. SysRanki and

P artRanki are the i-th ranked alternative by our system and the i-th ranked
alternative by participants5 , respectively.
We compared our method with a baseline. It is a method without the forget-
ting function, namely the Eq. (5) using the Eq. (1) instead of the Eq. (4). Table 2
shows the experimental result. Our method barely improved the RMSE, as com-
pared with the baseline. However, the difference was not statistically significant.
Alternatives in the low ranks were not essentially important to evaluate the
methods. Therefore, we also evaluated these methods with the top 3 alternatives.
3
agree(SysRanki , P artRanki )
T op3Acc = i (7)
3
where SysRanki and P artRanki are the same as the RMSE calculation. The
function agree(SysRanki , P artRanki ) is 1 if there is the agreement between
5
This is based on the average ranks among participants.
14 R. Kirikihira and K. Shimada

Table 3. The experimental result: accuracy of top 3.

Method D1 D2 D3 D4 Ave
Baseline 0.000 0.333 0.667 0.333 0.333
Proposed 1.000 0.333 0.667 1.000 0.750

SysRanki and P artRanki . Table 3 shows the result. The proposed method
obtained higher accuracy rates than the baseline on average. However, the value
of McNemar’s test was p = 0.125. This result does not always show a strong dif-
ference in correlation. We confirmed the limited success of our method with the
forgetting function. We need to evaluate our method with a larger test dataset.

4.2 Evaluation of DMA

We evaluated the effectiveness of our discussion map system with an assistant
function (DMA) described in Sect. 3.3 with four groups (G1, G2, G3, and G4).
Participants in each group are university students and not related to this work.
Each group discussed two topics (T3 and T4)6 .
(T3) Decide one product that boosts sales of a convenience store. As a condi-
tion, the target convenience store is a place where all participants know.
(T4) Decide one prefecture that you want to travel along with a foreign stu-
dent. As conditions, it is a trip of three days and two nights and there is no
limitation for traveling expenses.
To evaluate the effectiveness of the presence or absence of our system with DMA,
each group discussed one topic with our system7 and another topic without our
system, namely normal discussion without any systems. To avoid the bias based
on the order of the presence or absence of our system in the experiment, each
group discussed the order expressed in Table 4.
Although there are many evaluation points for the system, we evaluated
the effectiveness of our system by the following points: (1) time until the final
decision, (2) a satisfaction level about the discussion and (3) a satisfaction level
about the final decision. The range of the satisfaction levels was 1 (bad) to 10
(good).
Table 5 shows the discussion time until the final decision for each group.
The total time said that discussions without our system got to a solution faster
than those with our system (24.5 vs 26.4 min). Therefore, it seems that our sys-
tem did not work well. However, the discussion time depends on the number
of alternatives. If the number of alternatives is larger, the time to discuss them
becomes essentially longer. Therefore, we calculated the average discussion time
6
As conditions for the final decision, each discussion needs more than four alternatives
and more than two criteria.
7
The assistant function, DMA, becomes active in five minutes although groups with
our system can use the DM system from the start.
 Discussion Map with an Assistant Function for Decision-Making 15

Table 4. The experimental settings.

GroupID 1st discussion 2nd discussion

G1 T3 without system T4 with system
G2 T4 without system T3 with system
G3 T4 with system T3 without system
G4 T3 with system T4 without system

Table 5. The experimental result: total time (minutes). The numbers in each paren-
thesis denote time per alternative.

GroupID Without With

G1 21 (3.00) 19 (2.11)
G2 24 (4.00) 33 (3.30)
G3 23 (2.56) 21 (3.00)
G4 30 (5.00) 33 (3.67)
Average 24.5 (3.64) 26.5 (3.02)

for alternatives (the numbers in each parenthesis in the table). From the table,
our system reduced the discussion time per alternative (3.64 vs. 3.02). In addi-
tion, our system contributed to the increase of the number of alternatives that
participants discussed8 . This is a good point for making the final decision that
participants want. Visualizing alternatives and criteria by using our discussion
map system led to the improvement of the decision-making environment.
Next, we discuss the satisfaction levels of the test subjects. Table 6 shows
the results; sat.discussion denotes the satisfaction level about the discussion and
sat.decision denotes the satisfaction level about the final decision. For the sat-
isfaction level about the proceedings and the contents of the discussion from

Table 6. The experimental result: the satisfaction levels about the discussion and the
final decision. The dagger denotes the significant difference on the T-test.

GroupID sat.discussion sat.decision

Without With Without With
G1 8.25 9.00 9.00 9.75
G2 6.50 4.75 8.25 8.75
G3 4.25 6.00 7.25 8.75
G4 7.00 5.50 6.75 7.25
Average 6.50 6.31 7.81 8.63†

8
For instance, for G1, the number of alternatives with our system was 9 (19/2.11)
while that without our system was 7 (21/3).
16 R. Kirikihira and K. Shimada

the test subjects (sat.discussion), there is no difference between discussions

with/without our system (6.50 vs. 6.31). The role of the current system was
to support the consensus-building process. Therefore, the current system did
not always contribute to the early stage of each discussion, namely stimulating
divergent thinking. In addition, DMA occasionally did not work well, as an assis-
tant. DMA generates feedback comments in the situation that there are some
alternatives with high scores or there are gaps between overall and individual
references. In other words, the system just outputs “Let’s keep up the discus-
sion”, in the situation that any conditions mentioned above are not satisfied. It
led to the decrease of the satisfaction level about the discussion, as a support
system. As a result, the system did not obtain higher scores in this evalua-
tion point, sat.discussion. On the other hand, we obtained a good result for the
satisfaction level about the final decision (sat.decision). There was significantly
different between situations with/without our system (p = 0.041 on the T-test).
The reason why the sat.decision score improved was that each participant can
understand the difference between own thoughts and the whole opinion via out-
puts from DMA. DMA contributed to lax sharing of opinions in a positive sense.
This result shows the effectiveness of our system with an assistant function for
the group discussion.
After the experiment, we took a survey from test subjects. The followings
are the positive and negative opinions:

Positive opinions:
– It helped to easily understand the own ranking and the whole ranking for an
alternative.
– DMA’s suggestion was sometimes effective to break the silence in the stagnant
discussion.
– Visualization was effective to be clear about own thoughts.
Negative opinions:
– Sometimes, I struggled to concentrate on the discussion because of operations
for the DM system.
– It was difficult to understand the map when the number of nodes became
larger.

We obtained the positive opinions for the visualization and the assistant function.
On the other hand, our system needs to improve the operability. This is one
important future work.

5 Conclusions
In this paper, we proposed a tool for supporting consensus-building in conversa-
tions with multiple participants. Our method estimated the ranking of alterna-
tives in each discussion by using each discussion map. We introduced a forgetting
function into the consensus estimation model. In the experiment, the method
obtained the limited success as compared with a naive score calculation.
 Discussion Map with an Assistant Function for Decision-Making 17

We also implemented some feedback functions to participants during discus-

sions; charts and sentences. Visualizing information about the current state in
the discussion was useful for the participants to make the decision. We need
to evaluate other evaluation points although we evaluated our system with
sat.decision and sat.discussion in this paper.
The input for estimating the consensus in the current system is the discus-
sion map on tablet terminals or laptop PCs. However, conversations contain
many characteristics; verbal and nonverbal information. We have studied sev-
eral aspects for discussion situation, such as top-view images [14] and facilitators’
behaviors [16]. Integrating this discussion map system with these characteristics,
namely a multi-modal interpretation approach, is one interesting future work.
In the experiment, the topics of the discussion were just a case study. Evalu-
ation in real PBL situations or in decision-making tasks is also important future
work. The current system handles Japanese. However, the techniques in the sys-
tem are easily scalable to other languages. The discussion map system itself is
essentially language-independent. The assistant function for the sentence genera-
tion is based on a template-based generation scheme. Therefore, we can generate
other language templates easily. Experiments with the multilingual system are
interesting future work.

Acknowledgment. This work was supported by JSPS KAKENHI Grant Number

17H01840.

References
1. Alonso, S., Herrera-Viedma, E., Cabrerizo, F.J., Chiclana, F., Herrera, F.: Visual-
izing consensus in group decision making situations. In: IEEE International Con-
ference on Fuzzy Systems, FUZZ-IEEE 2007, pp. 1–6 (2007)
2. Bautista, J., Carenini, G.: An integrated task-based framework for the design and
evaluation of visualizations to support preferential choice. In: Proceedings of AVI
2006, pp. 217–224 (2006)
3. Ebbinghaus, H.: Memory: A Contribution to Experimental Psychology. Dover Pub-
lications, New York (1885)
4. El-Assady, M., Hautli-Janisz, A., Gold, V., Butt, M., Holzinger, K., Keim, D.:
Interactive visual analysis of transcribed multi-party discourse. In: Proceedings of
ACL 2017, System Demonstrations, pp. 49–54 (2017)
5. Gratzl, S., Lex, A., Gehlenborg, N., Pfister, H., Streit, M.: LineUp: visual analysis
of multi-attribute rankings. IEEE Trans. Vis. Comput. Graph. 19(12), 2277–2286
(2013)
6. Hmelo-Silver, C.E.: Problem-based learning: what and how do students learn?
Educ. Psychol. Rev. 16, 235–266 (2004)
7. Ito, T., Imi, Y., Ito, T., Hideshima, E.: COLLAGREE: a facilitator-mediated large-
scale consensus support system. In: Proceedings of the 2nd Collective Intelligence
Conference (2014)
8. Ito, T., Imi, Y., Sato, M., Ito, T., Hideshima, E.: Incentive mechanism for managing
large-scale internet-based discussions on COLLAGREE. In: Proceedings of the 3rd
Collective Intelligence Conference (2015)
18 R. Kirikihira and K. Shimada

9. Katsura, Y., Okada, S., Nitta, K.: Dynamic argumentation support tool using
argument diagram. In: Proceedings of The 29th Annual Conference of the Japanese
Society for Artificial Intelligence (2015). (in Japanese)
10. Masukawa, H.: Development of the reflective collaboration note: ReCoNote. In:
Proceedings of the 29th Annual Conference of JSET (2013). (in Japanese)
11. Miyake, N., Shirouzu, H.: The dynamic jigsaw: repeated explanation support for
collaborative learning of cognitive science. In: The Meeting of the 27th Annual
Meeting of the Cognitive Science Society (2005)
12. Nagao, K.: Meeting analytics: creative activity support based on knowledge discov-
ery from discussions. In: Proceedings of the 51st Hawaii International Conference
on System Sciences, pp. 820–829 (2018)
13. Nagao, K., Kaji, K., Yamamoto, D., Tomobe, H.: Discussion mining: annotation-
based knowledge discovery from real world activities. In: Aizawa, K., Nakamura,
Y., Satoh, S. (eds.) PCM 2004. LNCS, vol. 3331, pp. 522–531. Springer, Heidelberg
(2004). https://doi.org/10.1007/978-3-540-30541-5 64
14. Sakaguchi, K., Shimada, K.: Cooperation level estimation of pair work using top-
view image. In: Kim, S., Jung, J.-W., Kubota, N. (eds.) Soft Computing in Intelli-
gent Control. AISC, vol. 272, pp. 77–87. Springer, Cham (2014). https://doi.org/
10.1007/978-3-319-05570-1 9
15. Scardamalia, M., Bransford, J., Kozma, B., Quellmalz, E.: New assessments and
environments for knowledge building. In: Griffin, P., McGaw, B., Care, E. (eds.)
Assessment and Teaching of 21st Century Skills, pp. 231–300. Springer, Dordrecht
(2012). https://doi.org/10.1007/978-94-007-2324-5 5
16. Shiota, T., Yamamura, T., Shimada, K.: Analysis of facilitators’ behaviors in multi-
party conversations for constructing a digital facilitator system. In: Proceedings of
the 10th International Conference on Collaboration Technologies (2018)
17. Suzuki, H., Funaoi, H., Kubota, Y.: Supporting “assemble & disperse” style col-
laborative learning using tablet terminals. Technical report of IEICE-ET2013-26,
pp. 41–46 (2013). (in Japanese)
18. Takagi, H., Shimada, K.: Understanding level estimation using discussion maps for
supporting consensus-building. Procedia Comput. Sci. 35, 786–793 (2014)
19. Villalon, J.J., Calvo, R.A.: Concept map mining: a definition and a framework for
its evaluation. In: IEEE/WIC/ACM International Conference on Web Intelligence
and Intelligent Agent Technology 2008, pp. 357–360 (2008)
20. Yamasaki, K., Fukuda, H., Hirashima, T., Funaoi, H.: Kit-build concept map and
its preliminary evaluation. In: Proceedings of The 18th International Conference
on Computers in Education, ICCE 2010, pp. 290–294 (2010)
 An Integrated Support System
for Disaster Prevention Map-Making
Using Town-Walk Information Gathering

Sojo Enokida1 , Takashi Yoshino1(B) , Taku Fukushima2 , Kenji Sugimoto2 ,

and Nobuyuki Egusa1
1
Wakayama University, Sakaedani 930, Wakayama-City, Japan
enokida.sojo@g.wakayama-u.jp, yoshino@sys.wakayama-u.ac.jp
2
Osaka Institute of Technology, Osaka 573-0196, Japan

Abstract. Throughout Japan, numerous disaster prevention maps have

been made over time using the method of town-walk. This method will
assist in disaster prevention and enable a greater understanding of the
disaster prevention field. The development of these maps increases disas-
ter awareness, improves the self-help ability of individuals, and increases
cooperation amongst local communities. Currently, there is a support
system for developing conventional disaster prevention maps; however,
no maps have been designed using the town-walk method. Therefore, in
this paper, we present an integrated support system for making disas-
ter prevention maps using the information gathered by the town-walk
method. We conducted a comparison experiment between the proposed
system and the conventional system (using paper maps). This exper-
iment demonstrated that a consistent support function would be effec-
tive in developing disaster prevention maps using the town-walk method.
We also confirmed that the proposed system improves disaster aware-
ness among individuals, understanding of the area, and the conventional
system.

Keywords: Disaster prevention map · Town-walk · WebGIS

Disaster prevention awareness

1 Introduction
The Great East Japan earthquake caused damage to the local government. This
earthquake paved way to the concept of “public help from the central and local
governments” in a wide-area disaster. It is often said that disaster-prevention
measures should ideally be a combination of public help from central and local
governments, mutual help from local communities, and every individuals’ self-
help.
The map-making using the town walking technique is held in various places
in Japan. Its purpose is to improve the awareness of disaster prevention among
the participating people and to understand the area where they live. It has
c Springer Nature Switzerland AG 2018
H. Egi et al. (Eds.): CollabTech 2018, LNCS 11000, pp. 19–34, 2018.
https://doi.org/10.1007/978-3-319-98743-9_2
Another random document with
no related content on Scribd:
 the 26th April, at noon, accompanied by his aide-de-camp,
Captain del Pilar, and Mr. Leyba, his private secretary. …

"Throughout the whole stay of General Aguinaldo in Singapore

the editor was kept fully informed daily of the progress of
affairs. Naturally, however, all statement of what occurred
has been withheld by us until what has been deemed the fitting
moment has arrived. The substance of the whole incident in its
relations to the recent course of affairs in the Philippines
has been very fully telegraphed by the editor both to New York
and London."

{594}

Mr. Pratt, the U. S. Consul-General at Singapore, had already,

under date of April 28, given his own official report of the
interview with General Aguinaldo, to the Department at
Washington, as follows:

"I have the honor to report that I sent you on the 27th
instant, and confirmed in my dispatch Number 211 of that date,
a telegram which deciphered read as follows. … 'General
Aguinaldo gone my instance Hongkong arrange with Dewey
co-operation insurgents Manila.
PRATT.'

"The facts are these: On the evening of Saturday the 23d

instant, I was confidentially informed of the arrival here,
incognito, of the supreme leader of the Philippine insurgents,
General Emilio Aguinaldo, by Mr. H. W. Bray, an English gentleman
of high standing, who, after fifteen years' residence as a
merchant and planter in the Philippines, had been compelled by
the disturbed condition of things resulting from Spanish
misrule to abandon his property and leave there, and from whom
I had previously obtained much valuable information for
Commodore Dewey regarding fortifications, coal deposits, etc.,
at different points in the islands. Being aware of the great
prestige of General Aguinaldo with the insurgents, and that no
one, either at home or abroad, could exert over them the same
influence and control that he could, I determined at once to
see him, and, at my request, a secret interview was
accordingly arranged for the following morning, Sunday, the
24th, in which, besides General Aguinaldo, were only present
the General's trusted advisers and Mr. Bray, who acted as
interpreter.

"At this interview, after learning from General Aguinaldo the

state of and object sought to be obtained by the present
insurrectionary movement, which, though absent from the
Philippines, he was still directing, I took it upon myself,
whilst explaining that I had no authority to speak for the
Government, to point out the danger of continuing independent
action at this stage; and, having convinced him of the
expediency of co-operating with our fleet, then at Hongkong,
and obtained the assurance of his willingness to proceed
thither and confer with Commodore Dewey to that end, should
the latter so desire, I telegraphed the Commodore the same day
as follows, through our consul-general at Hongkong:
'Aguinaldo, insurgent leader, here. Will come Hongkong arrange
with Commodore for general cooperation insurgents Manila if
desired. Telegraph.
PRATT.'

"The Commodore's reply reading thus:

'Tell Aguinaldo come soon as possible.
DEWEY.'

"I received it late that night, and at once communicated to

General Aguinaldo, who, with his aide-de-camp and private
secretary, all under assumed names, I succeeded in getting off
by the British steamer 'Malacca,' which left here on Tuesday,
the 26th. Just previous to his departure, I had a second and
last interview with General Aguinaldo, the particulars of
which I shall give you by next mail. The general impressed me
as a man of intelligence, ability, and courage, and worthy the
confidence that had been placed in him.

"I think that in arranging for his direct cooperation with

the commander of our forces, I have prevented possible
conflict of action and facilitated the work of occupying and
administering the Philippines. If this course of mine meets
with the Government's approval, as I trust it may, I shall be
fully satisfied; to Mr. Bray, however, I consider there is due
some special recognition for most valuable services rendered.
How that recognition can best be made I leave to you to
decide."

Two days later (April 30), Mr. Pratt reported further, as

follows: "Referring to my dispatch No. 212, of the 28th
instant, I have the honor to report that in the second and
last interview I had with General Emilio Aguinaldo, on the eve
of his departure for Hongkong, I enjoined upon him the
necessity, under Commodore Dewey's direction, of exerting
absolute control over his forces in the Philippines, as no
excesses on their part would be tolerated by the American
Government, the President having declared that the present
hostilities with Spain were to be carried on in strict accord
with modern principles of civilized warfare. To this General
Aguinaldo fully assented, assuring me that he intended and was
perfectly able, once on the field, to hold his followers, the
insurgents, in check and lead them as our commander should
direct. The general further stated that he hoped the United
States would assume protection of the Philippines for at least
long enough to allow the inhabitants to establish a government
of their own, in the organization of which he would desire
American advise and assistance. These questions I told him I
had no authority to discuss."

Of the arrival of Aguinaldo at Hongkong and his conveyance

thence to Manila, the following account was given by Mr.
Wildman, the U. S. Consul at Hongkong, in a communication to
 the State Department at Washington, which bears date July 18:
"On May 2 Aguinaldo arrived in Hongkong and immediately called
on me. It was May 16 before I could obtain permission from
Admiral Dewey to allow Aguinaldo to go by the United States
ship 'McCulloch,' and I put him aboard in the night so as to
save any complications with the local Government. Immediately
on the arrival of Aguinaldo at Cavite he issued a
proclamation, which I had outlined for him before he left,
forbidding pillage, and making it a criminal offense to
maltreat neutrals. He, of course, organized a government of
which he was dictator, an absolutely necessary step if he
hoped to maintain control over the natives, and from that date
until the present time he has been uninterruptedly successful
in the field and dignified and just as the head of his
government. According to his own statements to me by letter,
he has been approached by both the Spaniards and the Germans,
and has had tempting offers made him by the Catholic Church.
He has been watched very closely by Admiral Dewey, Consul
Williams, and his own junta here in Hongkong, and nothing of
moment has occurred which would lead anyone to believe that he
was not carrying out to the letter the promises made to me in
this consulate. The insurgents are fighting for freedom from
the Spanish rule, and rely upon the well-known sense of
justice that controls all the actions of our Government as to
their future."

{595}

In reply to Consul Pratt's report of his interviews with

General Aguinaldo, and of his proceedings in connection with
the departure of that personage from Singapore to Hongkong,
the United States Secretary of State, Mr. Day, wrote, June 16,
as follows: "The Department observes that you informed General
Aguinaldo that you had no authority to speak for the United
States; and, in the absence of the fuller report which you
promise, it is assumed that you did not attempt to commit this
Government to any alliance with the Philippine insurgents. To
 obtain the unconditional personal assistance of General
Aguinaldo in the expedition to Manila was proper, if in so
doing he was not induced to form hopes which it might not be
practicable to gratify. This Government has known the
Philippine insurgents only as discontented and rebellious
subjects of Spain, and is not acquainted with their purposes.
While their contest with that power has been a matter of
public notoriety, they have neither asked nor received from
this Government any recognition. The United States, in
entering upon the occupation of the islands, as the result of
its military operations in that quarter, will do so in the
exercise of the rights which the state of war confers, and
will expect from the inhabitants, without regard to their
former attitude toward the Spanish Government, that obedience
which will be lawfully due from them. If, in the course of
your conferences with General Aguinaldo, you acted upon the
assumption that this government would cooperate with him for
the furtherance of any plan of his own, or that, in accepting
his cooperation, it would consider itself pledged to recognize
any political claims which he may put forward, your action was
unauthorized and cannot be approved."

Treaty of Peace, and Accompanying Papers

(55th Congress, 3d Session, Senate Document Number 62,
part 2, pages 337-354).

See, also (in this volume),

PHILIPPINE ISLANDS: A. D. 1806-1808.

UNITED STATES OF AMERICA: A. D. 1898 (April-June).

The War with Spain.
Movements or the Spanish squadron under Admiral Cervera,
and the blockading of it in the harbor of Santiago de Cuba.
Lieutenant Hobson's exploit.
The sinking of the collier "Merrimac" in the channel.

The opening of hostilities found a Spanish squadron of four

armored cruisers (the "Cristobal Colon," the "Almirante
Oquendo," the "Vizcaya," and the "Infanta Maria Teresa,") with
three torpedo-boat destroyers (the "Pluton," "Furor" and
"Terror") and some lighter craft, assembled at the Cape Verde
islands, under Rear-Admiral Pascual Cervera. They were in
Portuguese waters, and Portugal, though friendly to Spain, was
forced to issue a proclamation of neutrality, on the 20th of
April, which required the Spanish fleet to depart. Some of the
vessels then returned to Spain; but the seven named above
sailed westward, and their destination became a mystery, very
exciting for some time to the American mind. They might
attempt to surprise some American coast city; they might
intercept the battle-ship "Oregon," then making her way from
the Pacific coast, by the long circuit around Cape Horn; they
might have some plan for breaking the Cuban blockade. Acting
on the latter conjecture, and surmising that Porto Rico would
be chosen for the Spanish naval base, Admiral Sampson moved in
that direction to seek them. He attacked the forts at San Juan
(May 12), and satisfied himself that no fleet was in the bay.

The truth was that Cervera was then just entering the
Caribbean Sea, considerably to the south of Sampson's search.
He touched at the French island of Martinique, and at the
Dutch island of Curaçoa, and then slipped across to Santiago
de Cuba, where he was to be overtaken by his fate. In the long
hill-sheltered bay, with a narrow entrance, which forms this
excellent Cuban harbor, the Spanish fleet was so hidden that
nearly a fortnight passed before its whereabouts could be
fully ascertained. It was not until May 20 that a blockade of
Santiago was established by a flying squadron of the American
fleet, under Commodore Schley, with certainty that the
squadron of Cervera was harbored there. On the 1st of June,
Admiral Sampson arrived on the scene, with a stronger naval
force, and took command. To attempt to force the narrow
entrance of the harbor, strongly fortified and thickly mined
as it was, and attack the Spanish fleet in the bay, was not
deemed practicable. The course resolved upon was to hold the
enemy fast in the shelter he had sought, until Santiago could
be taken, by a land attack. In pursuance of this plan, an
exploit of splendid daring was performed, in the early morning
of June 3, by a young officer, Lieutenant Richmond Pearson
Hobson, with a crew of seven volunteers, who placed and sank a
huge coaling ship, the "Merrimac," in the channel that leads
into Santiago Bay. The following is Admiral Sampson's report
of the undertaking and its achievement:

"Before coming here, I decided to make the harbor entrance

secure against the possibility of egress of the Spanish ships
by obstructing the narrow part of the entrance by sinking a
collier at that point. Upon calling upon Mr. Hobson for his
professional opinion as to a sure method of sinking the ship,
he manifested a most lively interest in the problem. After
several days' consideration he presented a solution which he
considered would insure the immediate sinking of the ship when
she had reached the desired point in the channel. This plan we
prepared for before we reached Santiago. This plan included
ten electric torpedoes on the outside of the ship, each of 78
pounds of gunpowder, sinking the ship partially before going
in, cutting the sea valves, and opening the cargo ports. The
plan contemplated a crew of only seven men and Mr. Hobson, who
begged that it might be intrusted to him. The anchor chains
were ranged upon deck for both the anchors, forward and aft,
the plan including the anchoring of the ship almost
automatically. As soon as I reached Santiago and had the
collier to work upon the details were commenced and diligently
prosecuted, hoping to complete them in one day, as the moon
and tide served best the first night after our arrival.
Notwithstanding every effort, the hour of 4 o'clock in the
morning arrived and the preparations were scarcely completed.
After a careful inspection of the final preparations I was
forced to relinquish the plan for that morning, as dawn was
breaking. Mr. Hobson begged to try it at all hazards.

"This morning proved more propitious, as a prompt start could

 be made. Nothing could have been more gallantly executed. We
waited impatiently after the firing by the Spaniards had
ceased. When they did not reappear from the harbor at 6
o'clock I feared they had all perished. A steam launch, which
had been sent in charge of Naval Cadet Powell to rescue the
men, appeared at this time, coming out under a persistent fire
from the batteries, but brought none of the crew. A careful
inspection of the harbor from this ship showed that the
'Merrimac' had been sunk in the channel somewhat farther in
than had been intended.
{596}
This afternoon the chief of staff of Admiral Cervera came out
under a flag of truce with a letter from the Admiral extolling
the bravery of the crew in an unusual manner. I can not myself
too earnestly express my appreciation of the conduct of Mr.
Hobson and his gallant crew. I venture to say that a more
brave and daring thing has not been done since Cushing blew up
the 'Albemarle.'" The sunken ship did not actually block the
channel; but that fact takes nothing from the gallantry of the
exploit. Why the intended spot in the channel was missed was
explained by Lieutenant Hobson in a statement which he
afterwards made: "When the 'Merrimac' poked her nose into the
channel," says the Lieutenant, "our troubles commenced. The
deadly silence was broken by the swash of a small boat
approaching us from the shore. I made her out to be a picket
boat. She ran close up under the stern of the 'Merrimac' and
fired several shots from what seemed to be 3 pounder guns. The
'Merrimac's' rudder was carried away by this fire. That is why
the collier was not sunk across the channel. We did not discover
the loss of the rudder until Murphy [the volunteer assigned to
that duty] had cast anchor. We then found that the 'Merrimac'
would not answer to the helm and were compelled to make the
best of the situation. … Submarine mines and torpedoes were
exploded all around us, adding to the excitement. The mines
did no damage, although we could hear the rumbling and feel
the ship tremble. We were running without lights, and only the
darkness saved us from utter destruction. When the ship was in
the desired position and we found that the rudder was gone, I
called the men on deck. While they were launching the
catamaran I touched off the explosives. At the same time two
torpedoes, fired by the 'Reina Mercedes,' struck the
'Merrimac' amidships. I can not say whether our own explosives
or the Spanish torpedoes did the work, but the 'Merrimac' was
lifted out of the water, and almost rent asunder."

What followed, in the experience of the crew, when their

vessel went down, is described as follows by Lieutenant
Hobson, in a narrative of "The Sinking of the Merrimac," which
he published at a later day: "The stricken vessel now reeled
to port. Some one said: 'She is going to turn over on us,
sir,' to which I replied: 'No; she will right herself in
sinking, and we shall be the last spot to go under.' The
firing suddenly ceased. The vessel lowered her head like a
faithful animal, proudly aware of its sacrifice, bowed below
the surface, and plunged forward. The stern rose and heeled
heavily; it stood for a moment, shuddering, then started
downward, righting as it went. A great rush of water came up
the gangway, seething and gurgling out of the deck. The mass
was whirling from right to left 'against the sun'; it seized
us and threw us against the bulwarks, then over the rail. Two
were swept forward as if by a momentary recession, and one was
carried down into a coal-bunker—luckless Kelly. In a moment,
however, with increased force, the water shot him up out of
the same hole and swept him among us. The bulwarks
disappeared. A sweeping vortex whirled above. We charged about
with casks, cans, and spars, the incomplete stripping having
left quantities on the deck. The life-preservers stood us in
good stead, preventing chests from being crushed, as well as
buoying us on the surface; for spars came end on like
battering-rams, and the sharp corners of tin cans struck us
heavily. … When we looked for the life-boat we found that it
had been carried away. The catamaran was the largest piece of
floating debris; we assembled about it. The line suspending it
from the cargo-boom held and anchored us to the ship, though
barely long enough to reach the surface, causing the raft to
turn over and set us scrambling as the line came taut.

"The firing had ceased. It was evident the enemy had not seen
us in the general mass of moving objects; but soon the tide
began to drift these away, and we were being left alone with
the catamaran. The men were directed to cling close in, bodies
below and only heads out, close under the edges, and were
directed not to speak above a whisper, for the destroyer was
near at hand, and boats were passing near. We mustered; all
were present, and direction was given to remain as we were
till further orders, for I was sure that in due time after
daylight a responsible officer would come out to reconnoiter.
It was evident that we could not swim against the tide to
reach the entrance. Moreover, the shores were lined with
troops, and the small boats were looking for victims that
might escape from the vessel. The only chance lay in remaining
undiscovered until the coming of the reconnoitering boat, to
which, perhaps, we might surrender without being fired on. …
The air was chilly and the water positively cold. In less than
five minutes our teeth were chattering; so loud, indeed, did
they chatter that it seemed the destroyer or the boats would
hear. … We remained there probably an hour."

At daylight a steam launch approached, and was hailed by

Lieutenant Hobson, who judged that there must be officers on
board to whom it would be safe to surrender. He was more than
right. The commander of the launch was Admiral Cervera, in
person, who took the nearly exhausted men from the water and
treated them with great kindness, admiring the bravery of
their exploit, and sending a flag of truce to Admiral Sampson
to announce their safety. They were taken aboard the 'Reina
Mercedes,' and, as prisoners of war, were confined at first in
Morro Castle, and afterwards in the city. It so happened that
they were locked in the Morro during a bombardment of the
Spanish coast defences and fleet by ten of our vessels on June
6th, when about 1,500 projectiles were fired; and much anxiety
 and indignation were expressed in this country in view of that
circumstance; but Mr. Ramsden, British consul at Santiago,
explained in a despatch that they were removed as soon as
lodgings could be prepared in the barracks—actually on June
7th. They were released on July 6th in exchange for prisoners
captured by our forces.

{597}

UNITED STATES OF AMERICA: A. D. 1898 (April-July).

War with Spain,
Destruction of the Spanish fleet in Manila Bay.
Despatches of Admiral Dewey.
His relations with Aguinaldo, the insurgent chief.
Arrival of American troops for the occupation of the city.

Commodore George Dewey, commanding the Asiatic Squadron, then

awaiting orders at Hongkong, received on the 25th of April the
following despatch by cable from the Secretary of the Navy: "War
has commenced between the United States and Spain. Proceed at
once to Philippine Islands. Commence operations at once,
particularly against the Spanish fleet. You must capture
vessels or destroy. Use utmost endeavors." On the sixth day
after receiving these orders (namely on May 1st), he was able
to report from Manila, by a telegram sent from Hongkong on the
7th: "The squadron arrived at Manila at daybreak this morning.
Immediately engaged enemy and destroyed the following Spanish
vessels: 'Reina Christina,' 'Castillia,' 'Don Antonio de
Biloa,' 'Don Juan de Austria,' 'Isla de Luzon,' 'Isla de
Cuba,' 'General Lezo,' 'Marques del Duaro,' 'El Curreo,'
'Velasco,' one transport, 'Isla de Mandano,' water battery at
Cavite. I shall destroy Cavite arsenal dispensatory. The
squadron is uninjured. Few men were slightly wounded. I
request the Department will send immediately from San
Francisco fast steamer with ammunition. The only means of
telegraphing is to the American consul at Hongkong."
In due time the post brought particulars of the action, in the
following report from Commodore Dewey, dated May 4: "The
squadron left Mirs Bay, [China] on April 27, immediately on
the arrival of Mr. O. F. Williams, United States consul at
Manila, who brought important information and who accompanies
the squadron. Arrived off Bolinao on the morning of April 30
and, finding no vessels there, proceeded down the coast and
arrived off the entrance to Manila Bay on the same afternoon.
The 'Boston' and 'Concord' were sent to reconnoiter Port
Subic, I having been informed that the enemy intended to take
position there. A thorough search of the port was made by the
'Boston' and 'Concord,' but the Spanish fleet was not found,
although, from a letter afterwards found in the arsenal (
inclosed with translation), it appears that it had been their
intention to go there. Entered the Boca Grande, or south
channel, at 11.30 p. m., steaming in column at distance at 8
knots. After half the squadron had passed, a battery on the
south side of the channel opened fire, none of the shots
taking effect. The 'Boston' and 'McCulloch' returned the fire.
The squadron proceeded across the bay at slow speed, and
arrived off Manila at daybreak, and was fired upon at 5.15 a.
m. by three batteries at Manila and two at Cavite and by the
Spanish fleet anchored in an approximately east and west line
across the mouth of Bakor Bay, with their left in shoal water
in Canacao Bay. The squadron then proceeded to the attack, the
flagship 'Olympia,' under my personal direction, leading,
followed at distance by the 'Baltimore,' 'Raleigh,' 'Petrel,'
'Concord,' and 'Boston,' in the order named, which formation
was maintained throughout the action. The squadron opened fire
at 5.41 a. m. While advancing to the attack, two mines were
exploded ahead of the flagship, too far to be effective. The
squadron maintained a continuous and precise fire at ranges
varying from 5,000 to 2,000 yards, countermarching in a line
approximately parallel to that of the Spanish fleet. The
enemy's fire was vigorous, but generally ineffective.

"Early in the engagement two launches put out toward the

'Olympia' with the apparent intention of using torpedoes. One
was sunk and the other disabled by our fire and beached before
an opportunity occurred to fire torpedoes. At 7 a. m. the Spanish
flagship 'Reina Christina' made a desperate attempt to leave
the line and come out to engage at short range, but was
received with such galling fire, the entire battery of the
'Olympia' being concentrated upon her, that she was barely
able to return to the shelter of the point. The fires started
in her by our shell at this time were not extinguished until
she sank. At 7.35 a. m., it having been erroneously reported
to me that only 15 rounds per gun remained for the 5-inch
rapid-fire battery, I ceased firing and withdrew the squadron
for consultation and a redistribution of ammunition, if
necessary. The three batteries at Manila had kept up a
continuous fire from the beginning of the engagement, which
fire was not returned by this squadron. The first of these
batteries was situated on the south mole head at the entrance
to the Pasig River, the second on the south bastion of the
walled city of Manila, and the third at Malate, about one-ha]f
mile farther south. At this point I sent a message to the
Governor-General to the effect that if the batteries did not
cease firing the city would be shelled. This had the effect of
silencing them.

"At 11.16 a. m., finding that the report of scarcity of

ammunition was incorrect, I returned with the squadron to the
attack. By this time the flagship and almost the entire
Spanish fleet were in flames, and at 12.30 p. m. the squadron
ceased firing, the batteries being silenced and the ships
sunk, burnt, and deserted. At 12.40 p. m. the squadron
returned and anchored off Manila, the 'Petrel' being left
behind to complete the destruction of the smaller gunboats,
which were behind the point of Cavite. This duty was performed
by Commander E. P. Wood in the most expeditious and complete
manner possible. The Spanish lost the following vessels:
Sunk—'Reina Christina,' 'Castillia,' 'Don Antonio de Ulloa.'
Burnt—'Don Juan de Austria,' 'Isla de Luzon,' 'Isla de Cuba,'
 'General Lezo,' 'Marques del Duaro,' 'El Correo,' 'Velasco,'
and 'Isla Mindanao,' (transport). Captured—'Rapido,' and
'Hercules' (tugs), and several small launches. I am unable to
obtain complete accounts of the enemy's killed and wounded,
but believe their loss to be very heavy. The 'Reina Christina'
alone had 150 killed, including the captain, and 90 wounded.

"I am happy to report that the damage done to the squadron

under my command was inconsiderable. There were none killed,
and only 7 men in the squadron very slightly wounded. As will
be seen by the reports of the commanding officers which are
herewith inclosed, several of the vessels were struck and even
penetrated, but the damage was of the slightest, and the
squadron is in as good condition now as before the battle. I
beg to state to the Department that I doubt if any commander
in chief, under similar circumstances, was ever served by more
loyal, efficient, and gallant captains than those of the
squadron now under my command. … On May 2, the day following
the engagement, the squadron again went to Cavite, where it
remains. A landing party was sent to destroy the guns and
magazines of the batteries there. … On the 3d the military
forces evacuated the Cavite Arsenal, which was taken
possession of by a landing party."

{598}

Promptly in response to this report of his victory, a joint

resolution of thanks to Commodore Dewey and his officers and
men, by the two Houses of Congress, was despatched to them,
with announcement to the former of his promotion to the rank
of rear-admiral. The admiral replied, on the 13th, from
Cavite, making due acknowledgments, and adding: "I am
maintaining strict blockade of Manila by sea, and believe
rebels are hemming in by land, although they are inactive and
making no demonstrations. Great scarcity of provisions in the
city. I believe the Spanish Governor-General will be obliged
to surrender soon. I can take Manila at any moment. To retain
possession and thus control Philippine Islands would require,
in my best judgment, a well equipped force of 5,000 men."

On the 20th he reported, further: "Aguinaldo, the rebel

commander in chief, was brought down by the 'McCulloch' [from
Hongkong]. Organizing forces near Cavite and may render
assistance that will be valuable." On the 27th of June, in
reply to inquiries from the Navy Department, he explained his
relations with Aguinaldo, as follows: "Aguinaldo, insurgent
leader, with thirteen of his staff, arrived May 19, by
permission, on 'Nanshan.' Established self Cavite, outside
arsenal, under the protection of our guns, and organized his
army. I have had several conferences with him, generally of a
personal nature. Consistently I have refrained from assisting
him in any way with the force under my command, and on several
occasions I have declined requests that I should do so,
telling him the squadron could not act until the arrival of
the United States troops. At the same time I have given him to
understand that I consider insurgents as friends, being
opposed to a common enemy. He has gone to attend a meeting of
insurgent leaders for the purpose of forming a civil
government. Aguinaldo has acted independently of the squadron,
but has kept me advised of his progress, which has been
wonderful. I have allowed to pass by water recruits, arms, and
ammunition, and to take such Spanish arms and ammunition from
the arsenal as he needed. Have advised frequently to conduct
the war humanely, which he has done invariably. My relations
with him are cordial, but I am not in his confidence. The
United States has not been bound in any way to assist
insurgents by any act or promises, and he is not, to my
knowledge, committed to assist us. I believe he expects to
capture Manila without my assistance, but doubt ability, they
not yet having many guns. In my opinion these people are far
superior in their intelligence and more capable of
self-government than the natives of Cuba, and I am familiar
with both races."
 Report of the Secretary of the Navy, 1898,
volume 2, pages 67-72 and 103.

On the 30th of June, troops sent from San Francisco, to the

number of 2,500 officers and men, commanded by General T. M.
Anderson, arrived in Manila Bay, to co-operate with the navy
in taking Manila and occupying the city, when taken. They were
followed by a second expeditionary force, under General F. V.
Greene, which arrived July 17, and by a third, July 25 and 31,
with which came General Merritt, commanding the corps and the
Department of the Pacific. General Merritt's army then
numbered nearly 11,000 men, and it was increased during the
next few weeks to more than 15,000.

Reports of the War Department, 1898,

volume 1, part 2, page 499.

An English officer, Major Younghusband, who visited Manila at

this time, remarked: "It may, perhaps, with some confidence be
prophesied that when the cold fit, which will in due course
follow the warmth of the present enthusiasm, falls on the
nation, America will discover that the true parting of the
ways was … in having allowed Admiral Dewey to do more than
defeat the Spanish fleet and exact a heavy indemnity from the
city of Manila before sailing away." It would seem to be more
true, however, to say that the parting of the ways was when a
military expedition was sent from San Francisco to Manila, to
be landed, for the capture of the city and for the occupation
of the islands. It is claimed with reason that Admiral Dewey
could not "sail away," after the destruction of the Spanish
ships, because he needed the harbor he had seized, his fleet
having lost most of the privileges it had formerly been using
in neutral ports, when it became the fleet of a belligerent
power. To retain possession of Manila Bay while it was needed
by the American fleet was clearly a measure connected
legitimately with the general conduct of the war against
Spain. But it is difficult to see that the landing of soldiers
 on the island of Luzon and the capture of the city of Manila
added anything to the security with which the Bay was held for
the purposes of Admiral Dewey's fleet, or that it contributed
at all to the weakening of Spain in the war, and to the rescue
of Cuba from Spanish misrule. For two months, from the first
day of May until the last day of June, before a soldier
arrived, and for six weeks longer, before Manila surrendered,
Admiral Dewey appears to have been as fully and as
conveniently in possession of all the advantages that
harborage there could give him, as he was after the Spanish
flag had been lowered in the city and on the island.
Therefore, the American conquest of the Philippines does not
readily connect itself with the war for the liberation of
Cuba, as a necessary part of it, but presents itself to the
mind as a somewhat supplementary enterprise, undertaken with
objects of its own.

UNITED STATES OF AMERICA: A. D. 1898 (May-August).

Conduct of English and German naval officers at Manila.

While Admiral Dewey was holding Manila Bay, before the taking
of the city, there were many rumors and exciting stories
afloat, of offensive behavior towards the American fleet by
commanders of German war ships that were sent to the scene. As
far as possible, the facts were officially suppressed, in
order to avoid a quarrel between the two countries, and no
authoritative account of what occurred can be found. But some
incidents obtained publicity which are probably true in the
main. The first unpleasant happening appears to have been the
arrival in Manila Bay of a German naval vessel, which steamed
in with entire disregard of the blockading fleet, as though
the port was its own. Thereupon Admiral Dewey sent a forcible
reminder to the captain that he was intruding upon a blockade,
by firing a shot across his bow, and ordering him to heave to.
The German captain, in a rage, is said to have called on the
commanding officer of a British squadron that was in the Bay,
for advice as to what he should do, and was told that he owed
 the American Admiral an apology for his violation of naval
etiquette, well settled for such circumstances as those
existing in Manila Bay. According to the story, the British
commander, Captain Sir Edward Chichester, himself on the best
of terms with Admiral Dewey, visited the latter, on behalf of
the German officer, and made the matter smooth.
{599}
But, either through indiscretion of his own, or because he had
instructions to interfere as much as possible with the
proceedings of the Americans, the German commander continued
to pursue an offensive course. According to report, be went so
far as to stop a movement which Aguinaldo (then a recognized
ally of the United States) was making, to take possession of a
certain island, and to capture some Spaniards who were on it.
This provoked Admiral Dewey to a demonstration against him so
threatening that he drew back in haste, and the island was
occupied.

According to all accounts, Admiral Dewey showed unsurpassed

wisdom and dignity in meeting and checking these offensive
proceedings without allowing them to become a cause of
international quarrel; and he was happily aided in doing so by
the hearty support of the British naval commander. According
to still another report of the time, a German admiral, who had
come upon the scene, meditated an interference to forbid the
bombarding of Manila, when the city was about to be attacked,
and, calling upon Sir Edward Chichester to ascertain what
action the latter would take, was significantly told, "That is
only known to Admiral Dewey and myself,"—which convinced him
that his project was not wise. An English writer has related,
with much satisfaction, that when Sir Edward's ship, the
"Immortalité," finally steamed out of Manila Bay, returning to
Hong Kong, "every ship in the American fleet manned her yards
and gave the British man-of-war three cheers as she passed
along; and she with the answering signal, 'thank you,' flying
at her mast-head, went on her way."
UNITED STATES OF AMERICA: A. D. 1898 (June).
Act creating the United States Industrial Commission.

An Act "authorizing the appointment of a non-partisan

Commission to collate information and to consider and
recommend legislation to meet the problems presented by labor,
agriculture, and capital," was passed by Congress and approved
by the President June 18, 1898. It provided:

"That a commission is hereby created, to be called the

'Industrial Commission,' to be composed as follows. Five
members of the Senate, to be appointed by the presiding
officer thereof; five members of the House of Representatives,
to be appointed by the Speaker, and nine other persons, who shall
fairly represent the different industries and employments, to
be appointed by the President, by and with the advice and
consent of the Senate. … That it shall be the duty of this
commission to investigate questions pertaining to immigration,
to labor, to agriculture, to manufacturing, and to business,
and to report to Congress and to suggest such legislation as
it may deem best upon these subjects. … That it shall furnish
such information and suggest such laws as may be made a basis
for uniform legislation by the various States of the Union, in
order to harmonize conflicting interests and to be equitable
to the laborer, the employer, the producer, and the consumer.
… That the commission shall give reasonable time for hearings,
if deemed necessary, and if necessary it may appoint a
subcommission or subcommissions of its own members to make
investigation in any part of the United States, and it shall
be allowed actual necessary expenses for the same. It shall
have the authority to send for persons and papers and to
administer oaths and affirmations. … That it may report from
time to time to the Congress of the United States, and shall
at the conclusion of its labors submit a final report."

The Commission thus contemplated was duly appointed by the

President, and organized by the election of Senator Kyle for
its chairman. For the scope and plan of its investigations a
committee on procedure made the following recommendations,
which were adopted by the Commission and which have been
followed in what it has done:

"The main work of the Commission may … be said to be to study

and compare existing laws bearing upon industrial conditions,
here and elsewhere, to ascertain by competent testimony
wherein they are deficient, defective, inoperative, or
oppressive, and to recommend such remedial statutes as will
tend not only to make the conditions of industry more uniform
as between the several States, but to remove such existing
sources or causes of discontent, inequality, and injustice as
can be reached and regulated through legislation. … In order
to secure satisfactory results, it appears to your committee
imperatively necessary that the work shall be confined
strictly to the main purpose, viz, of ascertaining the nature
and effects of existing legislation, and the nature of
remedial legislation which may be necessary or desirable to
equalize conditions in industry and to remove any just grounds
of complaint on the part of either labor or capital or of the
people at large.

"To facilitate the progress of the work we recommend the

division of the Commission into four subcommissions of five
members each, to be severally charged with the investigation
of present conditions and the formulation of remedial
suggestions in the following branches of industry:

1. On agriculture and agricultural labor.

2. On the conditions of labor and capital employed in

manufacturing and general business.

3. On the conditions of labor and capital employed in mining.

4. On transportation.