An Ontology for Designing a Collaborative Platform Involving a

Multi-touch and Multi-user Interactive Table

eline Joiron

, Fr

eric F

urst

, Gilles Kassel

, Alistair Jones

, Jean-Paul A. Barth

Claude Moulin

and Dominique Lenne

MIS Laboratory, University of Picardy Jules Verne, 33 rue Saint Leu, 80039 Amiens Cedex 1, France

Heudiasyc, JRU CNRS 7253, Universit

e de Technologie de Compi

egne,

Centre de Recherches de Royallieu, BP 20529, 60205 Compiegne Cedex, France

Keywords:

Knowledge Engineering, Ontology Development, Collaborative Work, Interactive Device.

Abstract:

We have developed a platform running on a large multi-touch and multi-users table. Its intended application

domain is the collaborative preliminary design phase of engineering projects. We also developed an ontology

describing the virtual objects of the platform and the projects on which users collaborate. In this paper we

ﬁrst present the originality of this ontology. It extends the DOLCE-CORE ontology in order to model the true

nature of the virtual objects that the users can manipulate on the device and the users’ actions. We then present

the different roles played by the ontology in the development of the platform. It serves as a reference of the

semantics for the designers of the collaborative platform and as a model of the environment including table,

participants, projects, etc. This model is mainly used by the intelligent agents of a multi-agent system whose

objective is to support the participants during a working session around the table.

1 INTRODUCTION

Collocated collaborative work environments have be-

come an increasingly popular ﬁeld of research. Their

application is often focused on performing collab-

orative and conceptual design work, such as brain-

storming or project planning. The ﬁrst goal of such

systems is to provide a method for saving the con-

tent of collaborative work and distributing the data

throughout the team. A secondary goal is to engender

communication and collaboration amongst the team.

The devices for capturing input data from the teams

are the same ones which inhibit the productivity of

group for socio-technological reason. For example,

several meeting systems have been proposed to fa-

cilitate brainstorming which places its users at dif-

ferent desktop workstations with users focusing pri-

marily on their screen. These systems diminish group

awareness and collaboration because the attention is

distributed throughout the system. Creativity and

productivity are negatively impacted (Hilliges et al.,

2007).

In response to the limitations introduced by the

traditional hardware repurposed for collocated group-

ware, researchers are capitalizing upon the increased

Figure 1: Devices.

availability of vertical and horizontal interactive sur-

faces to fashion interactive whiteboards and table-

tops capable of supporting a social framework more

advantageous for group collaboration (Geyer et al.,

2011). The TATIN-PIC project at the University

of Technology of Compi

egne continues with this

approach in order to build a collaborative multi-

display working environment for design and engi-

neering teams. The hardware is composed of a large

156

Joiron C., Fürst F., Kassel G., Jones A., Barthès J., Moulin C. and Lenne D..

An Ontology for Designing a Collaborative Platform Involving a Multi-touch and Multi-user Interactive Table.

DOI: 10.5220/0004539401560163

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2013), pages 156-163

ISBN: 978-989-8565-81-5

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

multi-user multi-touch table, a large vertical white-

board, individual graphics tablets and interactive vo-

cal devices. The middleware is a multi-agent sys-

tem which provides ample extensibility and stability

to programmers when building applications for this

interactive environment. This multi-agent architec-

ture allows for easy integration of voice-controlled

personal assistant agents which can perform tasks for

the users around the interactive tabletop. Integrating

these personal assistant agents, is a matter of design-

ing a protocol for communication with these agents

with the middleware.

To help designing the TATIN-PIC platform (Jones

et al., 2011; Moulin et al., 2011), we deﬁned an on-

tology hereafter referred as OntoTATIN-PIC. It is pri-

marily aimed at helping the design team composed

of multidisciplinary researchers. It is provided as

a means to reduce the conceptual complexity of the

TATIN-PIC project. Indeed, one must account both

for the design projects done collaboratively by the de-

sign team, and for the platform itself. Concepts re-

lated to the project include project phases, individ-

ual and collective actions, designed artefact, used re-

sources, or team members’ roles. Concepts related

to the platform include all newly created objects, and

all actions allowed by the system. A second objec-

tive of OntoTATIN-PIC is to provide software agents

with the conceptual resources allowing them to build

a model of the on-going projects, which increases

their supporting capacities.

Because the ontology spreads across a vast do-

main, we propose to build it using a multi-level and

multi-component approach. Thus, OntoTATIN-PIC is

structured inline with the well known DOLCE-CORE

ontology in its more recent version (Borgo and Ma-

solo, 2009; Masolo, 2010), adding modules to de-

scribe sub-domains like actions or documents. To let

OntoTATIN-PIC answer the requirements, two ver-

sions of it have been designed using the OntoSpec

method (Kassel, 2005): (i) an informal speciﬁcation

using the formalism of the OntoSpec method, used as

a semantic reference for the project team; and (ii) an

implementation using the MOSS formalism (MOSS,

2000), used by the software agents of the platform.

The purpose of this paper is to present and discuss

the content of the ontology and its purposes. Sec-

tion 2 summarizes the ontological framework; Sec-

tion 3 describes some resources used for deﬁning

OntoTATIN-PIC, focusing on virtual objects appear-

ing on the graphics surface and on the actions needed

to manipulate them; Section 4 presents how the on-

tology is exploited; the conclusion presents ongoing

work meant to complement the ontology and improve

supporting features.

2 ONTOLOGICAL FRAMEWORK

We selected DOLCE-CORE (Borgo and Masolo,

2009; Masolo, 2010), to structure OntoTATIN-PIC.

This section summarizes the structuring principles of

DOLCE-CORE and presents two extensions required

to deﬁne OntoTATIN-PIC: (i) an extension to the do-

main of actions; and (ii) an extension to the domain

of entities bearing information.

2.1 Main Categories

DOLCE-CORE distinguishes four main categories of

concrete entities having spatio-temporal extensions

(Fig. 2).

Objects and Events are distinguished according to

the way they are located in space and time: Objects,

in particular physical objects like a person or a smart-

phone, are mainly linked to space, whilst Events are

mainly linked to time. Objects obtain a temporal posi-

tion through Events to which they participate. In par-

ticular Objects participate in their life, an Event with a

temporal extension. Conversely, Events obtain a spa-

tial location through Objects that generate them, e.g.

the physical space your body occupies corresponds to

the spatial extension of your reading of this article.

Qualities are dimensions or aspects according to

which agents perceive, classify and compare Objects

and Events. Individual Objects and Events possess

their own individual Qualities. Such individual Quali-

ties represent dimensions shared by classes of entities.

One should note (Fig. 2) that Object Qualities are dis-

tinct from Event Qualities: physical Objects have a

mass, have a shape, a texture, etc., whilst Events have

a duration, may be fast or slow, may overlap in time or

on the contrary happen before or after other Events.

Finally, Individual Qualities have a temporary lo-

cation in a space speciﬁc to their Quality kind (e.g.

”being a 2 meter length”, ”being a 36 second dura-

tion”).

Entities that we have just mentioned are linked

together through relationships. We mention here

two main relationships: Objects (resp. Events) have

for parts (hasForPart) other Objects (resp. Events);

Objects temporarily participate in (participatesIn)

Events.

Let us mention an important difference (in partic-

ular for OntoTATIN-PIC) among Objects. DOLCE

makes the distinction between Physical Objects hav-

ing a physical reality and Non-physical Objects hav-

ing a cognitive or social reality (Masolo et al.,

2004). Non-physical Objects exist because agents de-

sign them and communicate on them. Among Non-

physical Objects, Mental Objects depend on an agent,

AnOntologyforDesigningaCollaborativePlatformInvolvingaMulti-touchandMulti-userInteractiveTable

157

e.g. a mnemo-technical process belonging to a per-

son, whilst Social Objects depend on several agents,

e.g. a commercial company, rules of a game. Non-

physical Objects do have a spatial position, but only

indirectly, contrary to that of Physical Objects. As we

will see in Section 2.3, such a distinction is crucial to

account for the nature of objects being handled on the

tabletop. But in the mean time we extend DOLCE-

CORE with a set of minimal concepts to be able to

conceptualize actions.

2.2 Actions

A central concept is that of Action. Following tra-

ditional approaches from the philosophy of action

(Searle, 1983) we deﬁne an Action as an intentional

process controlled by an entity capable of intentions

(i.e. processes referring to the world in which the ac-

tion happens). Such Actions (Fig. 2) contrast with

Happenings that have no intentional cause. In the

philosophy of action, one usually distinguishes sev-

eral categories of intentions, which leads to distin-

guish several categories of actions according to the

nature of the control(Pacherie, 2008). One can thus

distinguish premeditated actions (initiated and con-

trolled by a prior intention), from automatic or routine

actions (controlled by an intention-in-action). Such

a dimension corresponds to the ﬁrst axis we select

to classify Actions, distinguishing between Deliberate

Actions and Non-deliberate Actions. We use a second

semantic axis, corresponding to the individual or col-

lective nature of the agent doing the Action. Thus, we

distinguish between Individual Actions and Collective

Actions, the latter having a collective agent, meaning

a set of agents having emerging properties like that of

being capable of action (Lawson, 2012). Within the

TATIN-PIC project, a design team is a Collective un-

dertaking a design project that is an example of Col-

lective Action.

Speciﬁc relationships allow to describe Actions.

In particular, an Action can be performed by the fact

of carrying another Action (isGeneratedBy)

; More-

over, to reﬂect the different ways in which entities

participate in Actions (e.g. an entity plays the role of

agent, another the role of instrument), we introduce

as many relations (called in the literature “casual re-

lations” or “thematic roles”) specializing the relation-

ship of participation (participatesIn) between Objects

and Events: hasForAgent, hasForInstrument, hasFor-

Result.

This generation can be causal in nature (e.g. I turn on

a lamp by pressing on the switch) or based on a social con-

vention (e.g. I indicate that I turn right by the fact to lift my

right arm).

Figure 2: Main generic categories of entities. A vertical link

indicates subsumption; an horizontal link indicates incom-

patible subsumed children.

2.3 Entities Bearing Information

A ﬁnal extension we introduced concerns the con-

ceptualization of textual or graphical inscriptions that

bear some meaning for agents. Such inscriptions are

related to the objects being handled on the multi-

touch table, e.g. a drawing representing a Post-it note,

a graphical representation of a project step. To de-

ﬁne OntoTATIN-PIC we reuse the I&DA ontologi-

cal module (Fortier and Kassel, 2004) that contains

a set of generic concepts concerning semiotics. Here,

we informally summarize the main concepts of I&DA

(Fig. 3).

Figure 3: Main concepts of I&DA.

A Conceptualization is a conceptual content.

I&DA distinguishes between two categories of Con-

ceptualizations, namely Concepts and Descriptions.

A Description loosely speaking, uses Concepts to re-

late to (refersTo) objects of the world. Among De-

scriptions, we distinguish between Propositions, cor-

responding to factual descriptions to which we can

assign a truth value, and Instructional Descriptions,

corresponding to imperative descriptions like ”do this,

then do that.”

An Expression is an expression which represents

the code of a Conceptualization in a communication

language, like the linguistic expression ”The room

temperature is high” expressing a Proposition. The

coding language may be text, graphics, or gestures. It

can be low-level not understandable by a human, but

rather interpretable by a machine.

KEOD2013-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

158

An Inscription is the material embodiment, by

means of stuff, of the shape of an entity. Creating an

Inscription can be done from an entity directly, like a

trace of steps or ﬁngerprints, or indirectly by embody-

ing an Expression referring to an entity, like a printed

image or text, or a Braille cell.

A Physical support is a physical object supporting

Inscriptions, that are temporarily stable in time. Some

Physical Supports play incidentally this role, like a

table being tagged, some dust in which something is

written. Other Physical Supports, known as Informa-

tion Media, are intentionally built to ﬁll this support

function and are thus technical artefacts, like a piece

of paper, a board, a memory space inside a computer.

3 OntoTATIN-PIC

Building OntoTATIN-PIC consisted ﬁrst in deﬁning

the various phases of the design process, namely,

Brainstorming, Causal Analysis or Functional Anal-

ysis. In order to describe them semantically, we ﬁrst

examined the various objects that were generated and

handled on the platform. The section focuses on the

deﬁnition and characterization of the different ob-

jects, especially in the combination table/interactive

board, and on the possible actions on such objects.

3.1 Manipulated Entities

TATIN-PIC device, composed of several media of

type Information Medium (an interactive tabletop and

an interactive board display which are coupled with

tablets and smartphones), allows participants to ma-

nipulate, through tactile interaction and / or voice,

light projections and act on entities that these lights

represent. Each participant thus has a tactile menu

and a tactile keyboard. Post-its notes are created and

grouped. Several diagrams can be constructed and ar-

ranged.

Light projections are Inscriptions, in the sense of

I&DA, more precisely Inscriptions of expression: they

materialize Expressions. From the point of view of

actions that the participants can complete, two main

categories of Inscriptions are to be distinguished: ﬁrst,

Inscriptions that pictorially represent physical objects

(e.g. a Post-it note) and permit on these Inscriptions

actions similar to those carried out on the physical ob-

jects; secondly, Inscriptions representing non physical

objects such as texts, groups of Post-its notes, tasks,

risks, etc. In this section, we present these two cate-

gories of Inscriptions.

3.1.1 Virtual Objects

Among all inscriptions we identiﬁed in OntoTATIN-

PIC, we deﬁned the concept of Virtual Object (Fig. 4).

This Inscription can on one hand represent a physical

object of the real world as a Picture (an Expression),

and on the other hand accept actions that can be done

on such an object. Take an apple for example. Its

projection onto the TATIN-PIC table is the image of

an apple (thus an expression). If it is possible to act

on this inscription, like peeling the apple or slicing

the apple, as could be done on the real object, then

that inscription becomes a ”Virtual Apple” instance

of the category Virtual Object as we have deﬁned it.

On the TATIN-PIC table a ﬁrst inscription having

this feature is the touch keyboard given to each partic-

ipant. It is a faithful representation of a standard key-

board with all the keys and is immediately received as

such by the participants. By using it, one can do the

same actions as with a real keyboard: type a letter,

erase text, validate commands, etc. We therefore in-

troduce the Keyboard Picture expression to represent

the picture of a keyboard, and the Virtual Keyboard

expression of the Virtual Object category.

Another Virtual Object that we have characterized

is the virtual Post-it note. Indeed, Post-it notes are an

essential part of brainstorming phases. Participants

use a large number or such notes to represent their

ideas, to organize them, progressively grouping them

before switching to other design phases. Post-it notes

on the TATIN-PIC tabletop appear as yellow rectan-

gles, like with standard paper Post-it notes, and offer

some of their options (only those that are relevant to

the application - for example it is not possible to tear a

virtual Post-it note on TATIN-PIC) permitting a triv-

ial and similar usage. Each participant can thus create

a new Post-it note, write down a text, move the note,

or else discard it. Thus, as for the virtual keyboard,

we distinguish the expression representing a Post-it

note picture (Post-it Picture) from the inscription of

the expression on the tabletop that can be handled by

a participant, i.e. the Virtual Post-it.

Figure 4: Virtual objects and inscriptions.

We reﬁne OntoTATIN-PIC, distinguishing several

types of Virtual Post-it. Thus, Blank Post-it represents

AnOntologyforDesigningaCollaborativePlatformInvolvingaMulti-touchandMulti-userInteractiveTable

159

an empty virtual Post-it note, i.e. a Post-it note bear-

ing no text. A Filled Post-it identiﬁes a virtual Post-it

note with a text. A Filled Post-it is an inscription that

has for part (hasForPart) another inscription that is

textual (Text inscription). This distinction allows us to

differentiate between the action of creating a Post-it

note (leading to a Blank Post-it), from that of ﬁlling a

Post-it note, transforming a Blank Post-it into a Filled

Post-it.

Likewise, in order to perform some actions, a par-

ticipant must ﬁrst identify a Post-it note on the table-

top, by selecting it, for example to discard it. This led

us to introduce the notion of Selected Post-it that is a

Virtual Post-it with the state ”selected”. In this case,

the color of the Post-it note on the table is modiﬁed

(the image changes). Conversely, an Unselected Post-

it represents any Virtual Post-it that is not selected.

As we mentioned previously, participants gather

Post-it notes together, making groups. These groups

are an opportunity to discuss in the next section, in-

scriptions of non physical objects called collections.

3.1.2 Collections

To conceptually represent the notion of group we

called upon the concept of Collection presented in

section 2. A Collection is a non physical object.

Participants do not manage a collection on the ta-

ble but its representation taking the shape of an im-

age on the tabletop. Thus, OntoTATIN-PIC, follow-

ing the I&AD principles, distinguishes the image of

the collection - an Expression called Collection Pic-

ture - a graphical representation of a Collection, from

the lighted inscription of that Expression, as projected

onto the tabletop and called Collection Inscription. A

Collection Inscription has for part (hasForPart) other

inscriptions: Virtual post-it or Collection Inscription.

Two types of expressions represent different ”views”

of the same collection: Expanded Collection Picture is

an image which allows viewing the entire content of

a collection, meaning all its elements and links. Con-

tracted Collection Picture is a compact view of that

collection that does not show its internal structure but

allows applying some actions to it.

As a non-physical object, a Collection exists only

when it is discussed. In other words, a collection

of Post-it notes (or of other collections) exists only

when the participants’ brainstorming evoke and rec-

ognize its existence. It could be the case without

the Collection being represented (by a Collection In-

scription). In this regard removing from the table a

collection inscription does not necessarily lead to the

removal of the collection (as a non physical object)

that it refers to. Finally, any action on an inscription,

has no de-facto effect on the collection itself. These

aspects will be developed in section 3.2 about possi-

ble actions on collections. The next section concludes

our presentation of objects evoking the contributions

of the triptych Inscription-Expression-referenced en-

tity in the project.

3.1.3 About Triptych

Inscription-Expression-Referenced Entity

As we have seen previously, OntoTATIN-PIC is based

on a triptych Inscription Expression and referenced

entity. This ontological framework can account for

different situations:

• A same Inscription materializes different Expres-

sions. This case occurs when, for example, a vir-

tual Post-it note color changes, which is equiva-

lent to materialize a different picture. This case

corresponds to a change of state of the Virtual

post-it, which retains its identity. A similar case

arises if participants have the possibility to cus-

tomize the appearance of the keyboard (Keyboard

picture change).

• A same entity is duplicated on one (or more) In-

formation medium. This case occurs, for exam-

ple, when a Virtual post-it or a Collection is copied

to another Information medium. Duplication is to

create a new Inscription with the same properties,

except the location. Subsequently, the conceptual-

ization can accompany different strategies: should

the changes on an Inscription (e.g. change of the

referenced entity, modifying only the Expression)

be reﬂected on to the Inscription of origin or not?

• A same couple (Inscription, Expression) changes

its referent. This case occurs, for example, when

a Virtual post-it comes to represent a task, a risk

in the project, or any other entity. This is a con-

venience that designers of the platform use, for

example, to represent a chronogram. Conceptu-

ally, this change is reﬂected in the reference link

(relationship RefersTo).

• A same entity is visualized using various Pictures.

This is particularly the case of a Collection assum-

ing an extended view (revealing images of its el-

ements) and a contracted view (corresponding to

the only presentation of its label).

To maintain such a model and act selectively on its

different components, e.g. to change the visual repre-

sentation of a Collection without changing its content,

participants can perform various actions. In the next

section, we present the conceptualization of these ac-

tions.

KEOD2013-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

160

3.2 Actions

Following the modeling of actions that we introduced

in Section 2, collective actions must be distinguished

from individual actions. Collective actions essentially

correspond to the different phases of the preliminary

design process, performed in a collaborative manner.

We focus here on the individual actions carried out by

the participants on the objects described in the pre-

vious section and place emphasis on two aspects of

our modeling of actions: i) precision in the identiﬁca-

tion of the objects on which actions bear and ii) usage

of different conceptual relationships to ﬁnely describe

actions.

On the ﬁrst point, it should be noted that actions

can bear both on Inscriptions (e.g., Moving a post-it,

Moving a collection), Expressions (e.g., Ro-sizing a

post-it, Expanding a collection) or other non-physical

objects (e.g., Adding a member to a collection). Thus,

for example, when a Collection is concerned, three

types of action are distinguished: those on the Col-

lection itself (aiming at changing its elements), those

dealing with its graphical representation (an extended

view integrating the presentation of its elements and

a contracted view are available) and those bearing on

its materialization on the table (the inscription can be

changed of place). Of course, actions on objects have

repercussions on other objects. For example, the addi-

tion of a Collection creates de facto a new Inscription.

These effects must be described because they partici-

pate in the semantics of the actions.

Concerning the second point, namely the level of

detail of the modeling of actions, we illustrate it on the

identiﬁed actions involving Virtual post-its (see Figure

5).

Each class of actions is ﬁrstly described by means

of casual relationships. Thus, Selecting a post-it is an

action that: hasForAgent a Participant; hasForData an

Unselected virtual post-it and hasForResult a Selected

virtual post-it (which is actually the same instance of

Virtual post-it, whose background color has been al-

tered as a side effect of the action, in the sense that this

is not an outcome intentionally targeted by the agent).

The instrument used to perform the action may also be

speciﬁed. Thus, Creating a ﬁlled post-it by voice is a

kind of action performed by means of an earphone in

which the participant gives the corresponding order.

In addition, several relationships between actions

are exploited:

• The relationship of specialization. For example,

Filling a post-it is a special case (specialization) of

Editing a post-it: its data is a Blank post-it, which

is ﬁlled with an editor. By contrast, Modifying

a post-it amounts to modifying the contents of a

Filled post-it.

• The relation of composition. For example, Cre-

ating a ﬁlled post-it is composed of two actions

in sequence: Creating a blank post-it and Filling

a post-it.

• The relationship isGeneratedBy. It allows to ex-

press that an action is consequence of another

(in particular that gestures or speech acts gener-

ate previously described actions). Thus, Adding a

member to a collection is a kind of action which

isGeneratedBy an action of the type Bringing to-

gether two inscriptions. Indeed, in order to add an

item to a Collection, one has to bring together on

the table the item and the Inscription of the Collec-

tion.

Figure 5: Individual actions concerning the manipulation of

Virtual post-it in the project TATIN-PIC.

Up to now, OntoTATIN-PIC does not contain the de-

scription of the gestures of the participants. For ex-

ample, Moving a post-it is performed by a gesture of

the hand which consists in pointing the ﬁnger towards

the Virtual post-it and moving it to its ﬁnal position. In

the future, we plan to complete OntoTATIN-PIC with

a conceptualization of these gestures in order to be

more accurate and complete in our reference seman-

tics.

4 ROLE OF OntoTATIN-PIC

We present in this section the different roles played

by OntoTATIN-PIC.

4.1 Reference Semantics

The ﬁrst role of OntoTATIN-PIC was to serve as a

common representation of the project for all members

of the team. The act of collaboratively building an on-

tology of the project gave the team the opportunity to

identify and deﬁne critical aspects of the system that

would have otherwise been left implicit and ambigu-

ous (e.g. the notions of project, phase and session).

AnOntologyforDesigningaCollaborativePlatformInvolvingaMulti-touchandMulti-userInteractiveTable

161

This act of building a common vocabulary is espe-

cially important for multi-disciplinary teams.

The ontology of possible user actions in the

TATIN-PIC system was also helpful to the design-

ers of the user interface of the system. Explicitly

specifying the requirements and outcomes of each ac-

tion allows the designers to perform a cognitive walk-

through of the entire system. For example, when a

user copies a Post-it note from one surface to another,

user interface designers much explicitly examine the

outcomes of such an action to decide whether future

modiﬁcations of this Post-it note should be propa-

gated throughout the system and also appear in its

copies. For another example, when an inscription of

a collection of Post-its notes is delete, the user inter-

face designers must determine what happens to the

collection, i.e. whether it is recorded somewhere as an

undoable action or if it simply ceases to exist. Such

questions often provoke helpful discussions on user

experience and intention.

4.2 Semantics fo Human-Machine

Dialog

In the TATIN-PIC system, users wear headsets and

can trigger speciﬁc actions through vocal commands

given to virtual personal assistants who can under-

stand natural language. These virtual personal assis-

tants are coded in OMAS (Barth

es, 2011). In this en-

vironment, each assistant can have its own ontology

for a semantic representation of the tasks and domain

in which it is operating. Such local ontologies are

derived from the general ontology of the TATIN-PIC

project and is implemented using a representation lan-

guage called MOSS.

MOSS is a complex frame-based representation

language, allowing us to describe concepts, individ-

uals, properties, classless objects, default values, vir-

tual concepts or properties. MOSS is centered on the

concept of property and adopts a descriptive (typical-

ity) rather than prescriptive approach, meaning that

defaults are privileged. MOSS has a query mecha-

nism, allows multilinguism, and has other features de-

tailed in its documentation (MOSS, 2000). Reasoning

is done via the query system. An example of imple-

mentation can be found in (Bettahar et al., 2009).

When a participant uses vocal commands, those

are analyzed with respect to the content of the ontol-

ogy and the description of the tasks to achieve, them-

selves described by concepts of the ontology. When

a task is selected, a dialog allows adding the neces-

sary information to perform the task, or if the task is

dangerous (e.g. deleting a collection) to ask for a con-

ﬁrmation before the action is completed.

4.3 Semantics for Exchanges among

Agents

OntoTATIN-PIC also helps deﬁning the protocol by

which the different software components of the plat-

form exchange information. Though the details of the

software architecture are outside the scope of this ar-

ticle, the components of this system (i.e. interactive

tabletop, whiteboard, headset) each have numerous

software processes responsible for sending and re-

ceiving information when a user performs an action.

The TATIN-PIC system is composed of two multi-

agent platforms having different agent communica-

tion languages (ACL). Transfer agents (gateways) re-

structure the messages from and to the other platform.

The exchange format follows the JSON grammar and

syntax and use performatives derived from FIPA, e.g.

inform, request, answer, cancel, or failure. The mes-

sage content is also a JSON object, the properties of

which correspond to concepts of the OntoTATIN-PIC

ontology and are parameters of the action to perform

by the agent receiving the message. The following

message is a request for creating a collection of Post-

it notes.

{"action" : "create",

"args" :

{"category" :

"collection",

"content" : "Structure",

"ref" : ["p3-2-1", "p3-2-2"]}}

Here, parameters depend on the nature of the ac-

tion requested. The content argument is the text to be

inserted in the Collection label and the ref argument

is the Post-it note list to be added to the collection

created.

5 CONCLUSIONS

In this paper we have presented the OntoTATIN-PIC

ontology used as a resource for the design of a collab-

orative platform involving a multi-touch and multi-

modal interactive work environment. This ontology

plays a fundamental role as a reference semantics for

the project team itself, but also for the human machine

dialog and the exchanges between middleware agents

of the platform.

OntoTATIN-PIC proposes a conceptualization of

collective and individual actions achieved by the par-

ticipants during collaborative activities and a concep-

tualization of objects created or produced during these

activities and objects that the platform allows users to

manipulate.

KEOD2013-InternationalConferenceonKnowledgeEngineeringandOntologyDevelopment

162

OntoTATIN-PIC is an extension of the founding

DOLCE-CORE ontology. This approach gives to the

necessary rigor for deﬁning the true nature of the dif-

ferent objects involved in actions. In particular, we

distinguish the actions having effects only on inscrip-

tions and the actions which can also have effects on

the expressions of these inscriptions or on the entities

they reference.

Today we have an informal speciﬁcation of our

ontology in the language of the OntoSpec methodol-

ogy and a partial implementation in the MOSS lan-

guage. Two research opportunities emerge: on a con-

ceptual level, it seems interesting to identify between

DOLCE and the applicative concepts of OntoTATIN-

PIC, an intermediate (and reusable) layer correspond-

ing to an ontology of virtual reality; on an operational

level, the issue of creating a detailed model of per-

formed actions (by populating the identiﬁed classes

with instances) is raised: it would help provide soft-

ware agents with a better understanding of the work

done.

ACKNOWLEDGEMENTS

The TATIN-PIC project is supported and funded by

the region of Picardy in France and the European

Union. Europe is engaged in the region Picardy

through FEDER

REFERENCES

Barth

es, J.-P. A. (2011). Omas - a ﬂexible multi-agent envi-

ronment for cscw. Future Generation Computer Sys-

tems, 27:78–87.

Bettahar, F., Moulin, C., and Barth

es, J.-P. A. (2009). To-

wards a semantic interoperability in an e-government

application. Electronic Journal of e-Government,

7(3):209–226.

Borgo, S. and Masolo, C. (2009). Foundational choices in

dolce. In Staab, S. and Studer, R., editors, Handbook

on Ontologies, International Handbooks on Informa-

tion Systems, pages 361–381. Springer Berlin Heidel-

berg.

Fortier, J.-Y. and Kassel, G. (2004). Managing knowledge at

the information level: an ontological approach. In In:

Proceedings of the ECAI’2004 Workshop on Knowl-

edge Management and Organizational Memories, Au-

gust 22 2004, pages 39–45.

Geyer, F., Pfeil, U., Hochtl, A., Budzinski, J., and Reiterer,

H. (2011). Designing reality-based interfaces for cre-

ative group work. In In: Proceedings of the 8th ACM

The content of this paper is the sole responsibility of the

authors and in no way represents the views of the Picardy

Region or the European Commission or their services.

Conference on Creativity and Cognition, ser. CC ’11,

pages 165–174, New York, NY, USA. ACM.

Hilliges, O., Terrenghi, L., Boring, S., Richter, H., and

Butz, A. (2007). Designing for collaborative creative

problem solving. In In: Proceedings of the 6th ACM

SIGCHI conference on Creativity cognition CC 07,

page 137.

Jones, A., Kendira, A., Lenne, D., Gidel, T., and Moulin, C.

(2011). The TATIN-PIC project: A multi-modal col-

laborative work environment for preliminary design.

In Computer Supported Cooperative Work in Design

(CSCWD), 2011 15th International Conference on,

pages 154–161.

Kassel, G. (2005). Integration of the dolce top-level

ontology into the ontospec methodology. CoRR,

abs/cs/0510050.

Lawson, T. (2012). Ontology and the study of social reality:

emergence, organisation, community, power, social

relations, corporations, artefacts and money. Cam-

bridge Journal of Economics, 36(2):345–385.

Masolo, C. (2010). Objects, events, qualities: An intro-

duction to formal ontological distinctions in dolce. In

Advanced course at ESSLII, Copenhagen, Denmark.

Masolo, C., Vieu, L., Bottazzi, E., Catenacci, C., Ferrario,

R., Gangemi, A., and Guarino, N. (2004). Social roles

and their descriptions. In Ninth International Confer-

ence on the Principles of Knowledge Representation

and Reasoning, pages 267–277. AAAI Press.

MOSS (2000). http//www.utc.fr/ barthes/MOSS. Universit

de Technologie de Compi

egne, France.

Moulin, C., Jones, A., Barth

es, J.-P., and Lenne, D. (2011).

Preliminary design on multi-touch surfaces managed

by multi-agent system. International Journal of En-

ergy, Information and Communications, 2(4).

Pacherie, E. (2008). The phenomenology of action: A con-

ceptual framework. Cognition, 107(1):179 – 217.

Searle, J. R. (1983). Intentionality: An Essay in the Philos-

ophy of Mind. Cambridge University Press.

AnOntologyforDesigningaCollaborativePlatformInvolvingaMulti-touchandMulti-userInteractiveTable

163