Professional Documents
Culture Documents
Conceptual Modeling 33rd International Conference ER 2014 Atlanta GA USA October 27 29 2014 Proceedings 1st Edition Eric Yu
Conceptual Modeling 33rd International Conference ER 2014 Atlanta GA USA October 27 29 2014 Proceedings 1st Edition Eric Yu
https://textbookfull.com/product/intelligent-virtual-agents-14th-
international-conference-iva-2014-boston-ma-usa-
august-27-29-2014-proceedings-1st-edition-timothy-bickmore/
https://textbookfull.com/product/mobile-web-information-
systems-11th-international-conference-mobiwis-2014-barcelona-
spain-august-27-29-2014-proceedings-1st-edition-irfan-awan/
https://textbookfull.com/product/combinatorial-optimization-and-
applications-12th-international-conference-cocoa-2018-atlanta-ga-
usa-december-15-17-2018-proceedings-donghyun-kim/
Conceptual Modeling
33rd International Conference, ER 2014
Atlanta, GA, USA, October 27–29, 2014
Proceedings
123
Lecture Notes in Computer Science 8824
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen
Editorial Board
David Hutchison
Lancaster University, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Alfred Kobsa
University of California, Irvine, CA, USA
Friedemann Mattern
ETH Zurich, Switzerland
John C. Mitchell
Stanford University, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
Oscar Nierstrasz
University of Bern, Switzerland
C. Pandu Rangan
Indian Institute of Technology, Madras, India
Bernhard Steffen
TU Dortmund University, 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, Saarbruecken, Germany
Eric Yu Gillian Dobbie Matthias Jarke
Sandeep Purao (Eds.)
Conceptual Modeling
33rd International Conference, ER 2014
Atlanta, GA, USA, October 27-29, 2014
Proceedings
13
Volume Editors
Eric Yu
University of Toronto
Faculty of Information
Toronto, ON, M5S 3G6, Canada
E-mail: eric.yu@utoronto.ca
Gillian Dobbie
University of Auckland
Department of Computer Science
Auckland 1142, New Zealand
E-mail: gill@cs.auckland.ac.nz
Matthias Jarke
RWTH Aachen University
Lehrstuhl Informatik 5
52056 Aachen, Germany
E-mail: jarke@dbis.rwth-aachen.de
Sandeep Purao
Penn State University
University Park
PA 16802, USA
E-mail: spurao@ist.psu.edu
Honorary Chairs
Stuart Madnick MIT, Boston, MA, USA
Colette Rolland Université Paris 1 Panthéon – Sorbonne, France
Conference Chair
Sandeep Purao Penn State University, State College, PA, USA
Workshop Chair
Marta Indulska University of Queensland, Australia
Publicity Chair
Matti Rossi Aalto University, Finland
Panel Chair
Sham Navathe Georgia Tech, USA
Tutorial Chair
Cecil Chua Auckland University, New Zealand
Sponsorships Co-chairs
Sandeep Purao Penn State University, USA
Aditya Ghose University of Wollongong, Australia
Wolfgang Maass Universität des Saarlandes, Germany
Finance Chair
Anna Squicciarini Penn State University, State College, USA
Webmaster
Jon Becker Penn State University, State College, USA
Program Committee
Jacky Akoka CNAM and TEM, France
Yuan An Drexel University, USA
Joao Araujo Universidade Nova de Lisboa, Portugal
Zhifeng Bao University of Singapore, Singapore
Sandro Bimonte IRSTEA, France
Shawn Bowers Gonzaga University, USA
Stephane Bressan National University of Singapore, Singapore
Stefano Ceri Politecnico di Milano, Italy
Roger Chiang University of Cincinnati, USA
Dickson Chiu University of Hong Kong, SAR China
Conference Organization IX
Additional Reviewers
Sponsors
Gold Sponsors
Silver Sponsors
Bronze Sponsors
Keynotes (Abstracts)
Data Science and Prediction
Vasant Dhar
Departmento de Informática
Pontifı́cia Universidade Católica do Rio de Janeiro (PUC-Rio)
Rio de Janeiro, Brazil
{furtado,casanova,simone}@inf.puc-rio.br
Giancarlo Guizzardi
Keynotes
A Semiotic Approach to Conceptual Modelling . . . . . . . . . . . . . . . . . . . . . . 1
Antonio L. Furtado, Marco A. Casanova, and
Simone Diniz Junqueira Barbosa
Unstructured Data
Schema-Independence in XML Keyword Search . . . . . . . . . . . . . . . . . . . . . . 71
Thuy Ngoc Le, Zhifeng Bao, and Tok Wang Ling
Database Design
A New Approach for N-ary Relationships in Object Databases . . . . . . . . . 209
Jie Hu, Liu Chen, Shuang Qiu, and Mengchi Liu
Departmento de Informática
Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio)
Rio de Janeiro, Brazil
{furtado,casanova,simone}@inf.puc-rio.br
1 Introduction
Our understanding of information systems comprises facts, events and agents. Every-
where the Entity-Relationship model is used. The existing entity instances and their
properties, i.e., their attributes and the relationships among them, are the facts that
characterize a state of the world. States are changed by the occurrence of events
caused by operations defined by pre-conditions and post-conditions that are in turn
expressed in terms of such facts. The event-producing operations are performed by
certain agents, in an attempt to satisfy their goals, once again expressed by facts. Ac-
cordingly, our specifications are divided into three schemas to introduce, respectively,
the classes of facts (static schema), events (dynamic schema) and agents (behavioural
schema).
It so happens that the pre-conditions to bring about an event may need to be first
fulfilled as part of the effect (post-conditions) of other events. This partial-order
dependence immediately suggests the recursive application of backward-chaining
plan-generators in order to find one or more sequences of operations (plans) able to
perform a transition from the current state to a state wherein the goals of an agent
hold. By using a logic programming notation to represent the three schemas and
2 Three-Schemata Specifications
We have been working with the conceptual modeling of information systems with a
database component, considering their static, dynamic and behavioral aspects.
The static aspect concerns what facts hold at some database state, conveniently
described in terms of the entity-relationship model.
The dynamic aspect corresponds to events that can produce state transitions.
Events result from the execution of operations, defined in a declarative style by
their pre-conditions and post-conditions, according to the STRIPS proposal [25].
A Semiotic Approach to Conceptual Modelling 3
The three aspects treated in the preceding section were integrated through the applica-
tion of a plan-recognition / plan-generation paradigm [35,38].
In order to make our conceptual specifications executable [32], we created an en-
vironment where entity and relationship classes, operations, and goal-inference rules
and agent profiles are all represented as Prolog clauses. Also written in Prolog, algo-
rithms were provided for planning and for the simulated execution of the generated
plans [17,18,19,29]. Moreover it was noted that simulation can become a useful re-
source to support learning or training [20].
The plan-recognition side of the paradigm is relevant, after the system has been
made operational, as a means to extend conventional query facilities towards truly
cooperative responses. Cooperation, as discussed in section 5, is most effective when
one can detect what the user is trying to accomplish. The plan-recognition algorithm,
which we adapted from [46], matches a few observed actions of the user against a
library of previously recorded typical plans. The observed actions can be taken from
the execution log, which is updated whenever each operation of a transaction of the
user's initiative is executed. As we explained in [39], the library of typical plans, in
turn, can be constructed by inspecting the log and extracting and filtering sequences
of executed operations, whereby the transition indicated in some goal-inference rule
has been achieved.
4 A.L. Furtado, M.A. Casanova, and S.D.J. Barbosa
5 Semiotic Completeness
frames, a data structure which would be of major importance for the practical applica-
tion of our ideas, as will be repeatedly stressed in the sequel.
Our paper presented at SBBD 2007 [2] can be regarded as a first attempt to deal
with paradigmatic relations in the context of databases. The motivating problem was
that databases, particularly when storing heterogeneous, sparse semi-structured data,
tend to provide incomplete information and information which is difficult to catego-
rize. The paper first considers how to classify entity instances as members of entity
classes organized in a lattice-like generalization/specialization hierarchy. Then, it
describes how the frame representation employed for instances and classes, as well as
the closeness criterion involved in the classification method, favors the practical use
of similarity and analogy, where similarity refers to instances within the same class,
and analogy involves different classes. Finally, the paper argues that similarity and
analogy facilitate querying semi-structured data.
A more in-depth investigation of classification methods based on frames was the
object of a more recent work [48]. In fact, the problem of data classification goes back
to the definition of taxonomies covering knowledge areas. With the advent of the
Web, the amount of data available increased several orders of magnitude, making
manual data classification impossible. The paper presents a tool to automatically clas-
sify semi-structured data, represented by frames, without any previous knowledge
about structured classes. The tool uses a variation of the K-Medoid algorithm and
organizes a set of frames into classes, structured as a strict hierarchy.
The next step, still focusing on paradigmatic relations and the corresponding
trope, metaphor, was to promote a reuse strategy, whereby new conceptual specifica-
tions might be partly derived from previous ones. A paper along this line was pre-
sented at CIKM [8]. Metaphor is not merely a rhetorical device, characteristic of
language alone, but rather a fundamental feature of the human conceptual system.
A metaphor is understood by finding an analogy mapping between two domains. The
paper argued that analogy mappings facilitate conceptual modeling by allowing the
designer to reinterpret fragments of familiar conceptual models in other contexts.
The contributions of the paper were expressed within the tradition of the ER model,
the Description Logic framework and as extensions of the OWL.
This reuse strategy was further examined in [7,30]. These papers argued in favor
of a database conceptual schema and Semantic Web ontology design discipline that
explores analogy mappings to reuse the structure and integrity constraints of concep-
tual models, stored in a repository. We presupposed that a team of expert conceptual
designers would build a standard repository of source conceptual models, which less
experienced designers would use to create new target conceptual models in other
domains. The target models will then borrow the structure and the integrity con-
straints from the source models by analogy. The concepts were once again expressed
in the contexts of Description Logic, the RDF model and OWL to reinforce the basic
principles and explore additional questions, such as the consistency of the target
model.
Reusing a conceptual schema is of course a multi-phase process. After finding a
suitable source schema, adaptations will often be needed in view of conflicts with the
target schema being designed. The notion of blending [24] was exploited for this
6 A.L. Furtado, M.A. Casanova, and S.D.J. Barbosa
6 Concluding Remarks
Edgewood Arsenal
The Ordnance Department, in making plans for a shell filling
plant, thought to interest existing chemical firms in the manufacture
of the required toxic materials. As plans developed, however,
difficulties arose in carrying out this program. The manufacture of
such material at private plants necessitated its shipment to the filling
plant at Edgewood. The transportation of large quantities of highly
toxic gases seemed attended with great danger. The Director
General of Railroads ruled that all such shipments must be made by
special train, a very expensive method of transportation. Still more
serious objections were encountered in the attempt to enlist the co-
operation of existing firms. They recognized that the manufacture of
such material would be attended by very great danger; that the work
would be limited to the duration of the war; and that the processes
involved, as well as the plants necessary for carrying out their
processes, would have little post-war value. Moreover, such firms as
had the personnel and equipment were already over-worked. With a
few exceptions (notably the American Synthetic Color Company, the
Oldbury Electro-Chemical Co., Zinsser & Co., and the Dow Chemical
Company) they were unwilling to undertake work of this character on
any terms whatever.
Early in December, 1917, therefore, it was decided to erect, on
the site of the shell filling plant, such chemical plants as would be
necessary to furnish the toxic materials required for filling the shell.
The Arsenal is situated in an isolated district, twenty miles east of
Baltimore, Maryland, on the Pennsylvania Railroad, and comprises
3,400 acres. Since the main line of the Pennsylvania Railroad runs
on one side of the tract, while on another is the Bush River, only a
few miles from its mouth in Chesapeake Bay, the tract was ideally
situated for shipping. This site was referred to, at first, as
“Gunpowder Reservation,” but on May 4, 1918, the name was
officially changed to “Edgewood Arsenal.”
Some idea of the extent of the work may be gained from the
following facts. On October 1, 1918, there were 233 officers, 6,948
enlisted men and 3,066 civilians engaged in work at Edgewood. 86
cantonments were built, accommodating about 8,500 men, while the
five officers’ barracks provided accommodations for 290. The
completed hospital unit consisted of 34 buildings, accommodating
420 patients under ordinary conditions. The total number of buildings
erected on the Arsenal grounds was 550. 14.8 miles of improved
roads were built, and 21 miles of standard gauge and 15 miles of
narrow gauge railway. A system furnishing 9.5 million gallons of salt
water and another furnishing two millions of fresh water daily were
successfully installed. Large power plants were built in connection
with the shell filling plants and the chlorine plant.
Plants for phosgene, chloropicrin, mustard gas, chlorine and
sulfur chloride were built and placed in successful operation. Most of
the raw materials, with the exception of sulfur chloride, were
obtained from commercial firms. The other gases and manufactured
materials used, such as phosphorus, tin and silicon tetrachlorides,
bromobenzylcyanide and arsenic derivatives were supplied by
various plants scattered through the East and Middle West States.
The raw materials used by the Arsenal in 1918 were as follows:
Salt 17,358,000 pounds
Bleach 42,384,000 “
Picric acid 3,718,000 “
Alcohol 3,718,000 “
Sulfur 24,912,000 “
Sulfur chloride 6,624,000 “
Bromine 238,000 “
Benzyl chloride 26,000 “
The production of toxic materials and the amount shipped
overseas in bulk follow:
Production, Shipped in
Bulk,
Pounds Pounds
Chlorine:
Liquid 5,446,000 2,976,000
Gaseous 2,208,000
Chloropicrin 5,552,000 3,806,000
Phosgene 3,233,070 840,000
Production, Shipped in
Bulk,
Pounds Pounds
Mustard gas 1,422,000 380,000
Bromobenzyl
10,000
cyanide
White phosphorus 2,012,000 342,000
Tin tetrachloride 2,012,000 212,000
Titanium
362,000
tetrachloride
For nearly a month previous to the signing of the Armistice, the
various plants at the Arsenal had shut down or were operated only to
an extent sufficient to maintain the machinery and equipment in good
working order, on account of the lack of shell into which to fill the
gas, so that the above figures do not at all represent maximum
productive capacity.
These plants will be described in the appropriate chapters.
The shell filling plant was really composed of several small
plants, each of which was made up of units radiating from a central
refrigeration plant which would serve all the units. Each unit could
then be fitted with machinery adapted for filling shell of a different
size, and for a particular gas. Moreover, an accident in one of the
units would in no way impair the working of the remainder.
The problem involved in the filling of a shell with toxic material
(which is always a liquid or a solid and never a gas under the
conditions in which it is loaded in the shell) is similar in a way to that
of filling bottles with carbonated water. In the development of plans
for the filling plant, many suggestions were obtained from a study of
the apparatus used in commercial bottling plants. It was necessary
to keep in mind not only the large number of shell to be filled, but
also the highly toxic character of the filling material to be used. It was
essential that the work of filling and closing the shell should be done
by machinery in so far as that was possible, and that the operation
should be carried out in a thoroughly ventilated room or tunnel,
arranged so that the machinery contained in the tunnel could be
operated from the outside. Special care was taken in closing the
shell, the closing being accomplished by motors actuated by
compressed air, which, in the closing process were driven until they
stalled. In this way a uniform closing torque was obtained. The final
results secured were admirable, as is evidenced by the fact, reported
by the Quartermaster Officer at Vincennes on November 15, 1918,
that not a single leaky shell had been found among the 200,000 shell
received up to that date.
75 mm. Shell
Shipped
Filled
Overseas
Phosgene 2,009
N. C. 427,771 300,000
Shipped
Filled
Overseas
Mustard gas 155,025 150,000
Livens Drum
Phosgene 25,689 18,600
Grenades
White phosphorus 440,153 224,984
Tin tetrachloride 363,776 175,080