Co operation Using Collective Personal Connections
{ Adaptable Menu-based Query Interface {
Hiroaki Ogata a
,Yoneo Yano a
,NobukoFurugori b
and Jin Qun b
a
Department of Intelligentsystems and Informationsicence, Faculty of Engineering,
TokushimaUniversity, 2-1 Minami Josanjima, Tokushima770, Japan
b
Advanced Research Center of INES Corporation, 3-9-2 Ushikubo,Kohoku-ku,
Yokohama223, Japan
E-mail: [email protected]
This pap er describes a groupware system, called PeCo-Mediator, and its adaptable
menu-basedquery user-interface(UI). PeCo-Mediator collectsgroup users'personalcon-
nections (PeCo) to help users nding partners who can solve their problem in business
activities. Moreover, its UI is adaptable for a user's original p erspective and another's
viewp oint touse e ectivelydiverse personal information.
1. INTRODUCTION
The Combined actions of two or more individuals realize net b ene ts that more than
doublethe b ene ts available toasingle individual. Human-humancooperationindicates
the importanceofcreating anenvironmentwhere re nedsolutionscan b ebasedonargu-
mentand the resolution ofdi ering viewpoints[1]. It is signi cantto nd amore capable
and co operative partnerto meet the goal of cooperation.
Ourresearchfocusesonhowausercan ndasuitablepartnerwho cansolveaproblem.
A \PeCo" (p ersonal connection) is often a starting point for nding vital partner(s)[2].
Users often nd suitable partners of their acquaintances by using diverse p ersonal infor-
mationsuchasspecialabilities,interestsandbackgroundinso cial interactions. Also,the
trust based onthe personalrelationships makes iteasy toobtain the cooperation.
We proposea groupware system[3] , called \PeCo-Mediator,"that allowsgroup mem-
b ers(users) tocollectand use their PeCos[4]. All the userso erthe information ontheir
acquaintances and the closeness of their relationships to the system. The system helps
users nd a vital partner who can assist their problem solving, and supports access to
the backgroundofthe hitherto unknownpartners. Tohandling diverse anddynamic p er-
sonal information whichusers obtained intheir social relationships, the system will need
the maximum exibility in storing data from every user's p erspective and the utmost
adaptability inusing data from every possible viewpoint.
Inthispaper,wedescribetheadaptabilityofthemenu-basedqueryuser-interface(MUI)
whichmakesthe most use of diverse p ersonal information. Thecharacteristicfeatures of
the interface lie its allowing the user: 1) to freely customize the menu structure from
Main W indow
M enu- based Q uery I nterf ac e
PeC os D isplay W indow
Close
A
B
user B
user's acquaintance
user's acquaintances
target person
target's acquaintance
Icons
target person
C ustomization W indow
Figure 1. Screenof PeCo-Mediator.
his/her viewpoint; 2) to adopt others' menu structure as his/her own; and 3) to help
users nd partnersnot onlyfromhis/herownviewpointbutalsofromothers'viewpoints.
Namely, the menu is adaptable for a user's originalviewpointor another's one in its use
to enhance human-computerinteraction.
2. OVERVIEW OF \PeCo-Mediator"
2.1. Prototyp e system
Wedevelop ed a prototyp e system on a workstation (see Fig. 1). The system consists
of several clients and a server program and they are connected via the Internet. When
theyhavesomeproblems,users ndtargetpartnersthrough theMUIinMainWindowby
usingpersonalinfomaion. PeCosDisplayWindowshowstheclosenessofthe relationships
between a target and his/her acquaintances. The shorter the distance between a target
smo othly cooperate with the target by the introductionof Mr. A and Mr. B.
2.2. Flexible database
Currently, the most popular type is based on the relational data mo del, where data
are stored in records with pre-de ned elds. To gather diverse and dynamic personal
information, itis hardertopre-plan itsdata structures and predictallthe ways inwhich
the data will b e lo oked at. PeCo-Mediator handles such information using TRIAS[5]
whichallowsusers toeasilyadd, deleteorchangeattributes orvalues atany timeduring
its use. TRIAS represents data with triplets by a small grain size as (entity, attribute,
value).
TRIAS can search triplet data employing search conditions where any elements can
b e replaced by \3"; (E;A;V), (E;3;3), (3;A;3), (3;3;V), (E;A;3), (3;A;V), (E;3;V),
(3;3;3). For example, users obtain \Ogata" and \Yano" whose address is Tokushima
by the condition (3, address, Tokushima), when TRIAS has data such as (\Ogata",
\address", \Tokushima"), (\Ogata", \interests", \programming"), (\Yano", \address",
\Tokushima"), and (\Yano",\position", \professor").
2.3. Menu-based query interface
people
organization
interests
name
address
abilities
sex
department
post
home town
alma mater cooking
chess
movie
program
baseball
sight seeing
video game
games
skiing
entity attribute value
Figure 2. Flowof MUI.
use
store
Data
MUI
user
update
Figure 3. Basic frame work of
MUI.
Several alternative typ es of interfaces have b een developed, including MUI, graphical
interfaces,naturalandpseudonaturallanguageinterfaces,andbrowsers. Theseinterfaces
areorientedtowardsnonprogrammersandnoviceusers. Aresearchvalidatedthatevena
noviceusercouldmostsmo othlyconstructdesirableanswersthroughaMUI[6]. Therefore,
we adopted the interface intoPeCo-Mediator.
The system generates the menu structure from triplets in the database and makes a
hierarchy of entity, attribute, and value fromthe contentsof the database. Fig. 2 shows
menustructureof\interests." Inthis gure,theuserretrievespeoplewhoareinterestedin
\chess", \games",or\videogame"byemploying(3,\interests",\chess"),(3,\interests",
users can nd partners without the familiarity with the contents or organization of the
database.
Fig. 3 showsthe simple framework of PeCo-Mediator. Users can store every p ersonal
information whichthey obtained in their social activities. The menu items are automat-
ically updated according to additions, deletions or changes of attributes or values in the
database. In other words, the menu has the autonomy to adopt the database contents
into itself. Therefore,users can browse the contentsof the database inthat time.
3. ADAPTABLE MENU FOR A USER'S VIEWPOINT
use
store
Data
Adaptable MUI
adapt
user
update
Figure 4. Frame work of
adaptabele MUI.
Groupwaresystemsareusedbyseveralworkersinagroup.
Users' expertise and interpretation of data are di erent[7].
Ordinary MUI which have the only prede ned contents are
short ofthe exibility. According toa user'srespectiveview-
point, PeCo-Mediator allows him/her to customize him/her
menu structure freely and to adapt their conceptual struc-
tures to the menu (see Fig. 4). By this feature, users can
get rid of needless information and reachneeded information
intuitively. That is one kindof information ltering systems.
Fig. 5 showsmenu structures of UserA and User B after
customizing respective menu such as Fig. 2. In this case,
User A divides into \outdoor" and \indoor," and User B is
interestedin \sports"and \game."
In this way, dummy values which the user inputed clarify the relationships between
the values (concepts). The viewof information throughout his/her own menu is sure to
enhance human-computer interaction between the user and the system. In Customiza-
tion Window, users graphically customize their respective menu structure by using the
command buttonssuch as \insert"(see Fig. 1).
4. ADAPTABLE MENU FOR DIFFERENT VIEWPOINTS
Theresearchoncustomizablesoftware(i.e.,X-windowcustomizing le)pointedoutthat
users requested to share their customization in its design. This study has demonstrated
thatcustomizationisnotsurelyindividualactivity. Membersofanorganizationmaya ect
eachother'sbehaviorforlongperiods. Consideringthereport,weproposethemulti-users
frame work, extendingthe single user one (see Fig. 6 ). Its main characteristicis that a
user obtains useful ideas through browsing di erentmenu structures.
4.1. Customization of menu structure
InPeCo-Mediator,auserselects thesuitablestructure fromthreecandidatessuchasa
nominateduser'smenuanddomainknowledgebases. PeCo-Mediatorhasthetwowaysto
tell auser some useful ideasfromthe di erentviewpoints. One is\bottom up approach"
and the other is \top down approach." Bottom up approach derives the user unknown
upper concepts from the user selected value. For example, when User A wants to know
others
sports
interests
cooking
chess
movie
skiing
program
baseball
sight seeing
games
video game
board game
Menu structure of User B
outdoor
interests
cooking
chess
movie
skiing
program
baseball
sight seeing
indoor
video game
game
true value
dummy value
NOT E
Menu structure of User A
Figure 5. Menustructures afterthe customization.
what \video game"is included by,the system shows\game"and \indoor" asthe answer
by the selection of \video game" and \upper button" in Fig. 7.
Ontheotherhands,intopdownapproach,UserA ndssomeunderconceptsof\indoor
" except the user's menu structure by the selection of \indoor" and \under button." In
Fig. 7, user A can adopt the idea(s) of User B into his menu. User A may nd an
advantageousoradaptedideatoclassify\indoor"under\games"and\others." Ofcourse,
this is realizedby the permission of UserB toUser A. Moreover,the system showsusers
the domain knowledgebases suchas address.
4.2. Finding a partner
When users retrieve partnersthrough users' menu,PeCo-Mediator allowsthem to use
di erent menus temporarily. Forexample, as shown in Fig. 7, User A can use User B's
menu instead of User A's menu in Fig. 5 by selecting User B's viewp oint from shared
viewp oints. Moreover,UserA can also use apart of thegeneral menu. Inthis way,users
can nd partners not onlyfrom users' originalviewpoint but also from others'ones.
5. CONCLUSION
Ourresearchfocusesonthewaysmemb erscan nd partnersforhuman-humanco oper-
ationusingtheirPeCos. Inthis pap er,weproposedanadaptableMUIwhichallowsusers
tosystematizepersonalinformationaccordingtotheirviewpoints. The questionnaireson
the MUI indicate that it is ecacy for most users, and that its adaptability is ecient
and signi cant forbrowsing through personalinformation to nd suitable partners.
REFERENCES
1. A. A. Clarke and G. G. Smyth: \Co-operative Computer based on Principles of
use
store
Data
adapt
user A
update
user B
update
use
store adapt
reference
Adaptable MUI
Adaptable MUI
Figure 6. Frame work of
Adaptable menu for dif-
ferent viewpoints.
outdoor
interests
skiing
baseball
sight seeing
indoor
others cooking
chess
movie
program
sight seeing
games
video game
Figure 7. UserA's menu structure afteradopting the idea of
UserB.
Human Co-operation", Int.J. Man-Machine Studies, No. 38, pp. 3-22, (1993).
2. K. Simomura: \Fifty- ve Points for Extending Personal Connections", PHP, p. 237,
(1988), inJapanese.
3. C. Ellis, J. S. Gibbs and L. G. Rein: \Groupware Some Issues and Exp eriences",
Commun. of ACM,Vol.34, No. 1,(1991).
4. H. Ogata, Y. Yano, N. Furugori, and Jin, Q.: \Mediator: Supp orting business ac-
tivities With Common exploitation of Personal Connections", Proc. of Japan-U.S.A
symp osium of Flexible Automation' 94, pp. 105-108, (1994).
5. Y.Yamamoto,A.Kashihara,K.Kawagishi,andN.Tsukamoto: \ATo olforConstruc-
tion of Personal Database : TRIAS", the trans. of IPSJ, Vol.30, No. 6, pp. 733-742,
(1989).
6. E. J. Bell and A. L. Rowe: \An Exploratory Study of Ad Hoc Query Languages to
Databases", Proc.of IEEE Eighth Int'lConf. onData Eng., pp. 606-613 (1992).
7. E. W. Mackey: \Patterns and SharingCustomizable Software",CSCW '90, pp. 209-
221 (1990).
