RESOURCE AGGREGATION IN DIGITAL LIBRARIES
Static vs Dynamic Protocols
Pedro Almeida, Marco Fernandes
IEETA, University of Aveiro, Aveiro, Portugal
Joaquim Arnaldo Martins, Joaquim Sousa Pinto
IEETA/DETI, University of Aveiro, Aveiro, Portugal
Keywords: Digital Libraries, Metadata Aggregation, Distributed Systems.
Abstract: This article analyses development of static and dynamic protocols to aggregate metadata and resources from
heterogeneous systems. In particular, it compares the advantages and drawbacks of both types of protocols
and presents a case study of the University of Aveiro Information System as an example of the possibilities
of dynamic resource aggregation systems.
1 INTRODUCTION
Resource harvesting has been a major issue in
Digital Libraries systems.
Gathering information from distinct sources with
different metadata schemas lead to the development
of distributed search mechanisms and resource
harvesting protocols. The main concept being
explored behind these approaches is the
development of interoperable and efficient federated
search engines, which can access, search and
retrieve metadata from different Digital Libraries
catalogues.
As the content of Digital Libraries evolves to
heterogeneous collections, composed by several
types of documents, new standards are being
developed and used to properly classify the
information contained in Digital Libraries. While
some years ago we could think only in text based
searches, the reality nowadays is that new formats of
queries are being developed, and the future and
success of huge repositories depends on the ability
to process text, image or sound based queries.
In order to provide federated search facilities over
metadata from different repositories, two different
strategies can be adopted: a static approach - harvest
the content of the repositories, index the metadata
and provide search mechanisms over it, or a real-
time dynamic approach - provide programmatic
interfaces to search and retrieve the metadata on the
fly.
Each one of these approaches has advantages and
drawbacks. In this article we analyze the
implications in using either a static or a real-time
dynamic aggregation protocol. In the end, we will
present a case study of a dynamic integration system
that was developed for the University of Aveiro's
Institutional Repository (IR), a Digital Library that
contains several documents of different formats
(text, images, videos and sounds).
2 STATIC AGGREGATION AND
DYNAMIC AGGREGATION
The aggregation of Digital Libraries resources is
concerned with two important issues of Information
System technologies: provide mechanisms to
simultaneously search several repositories of
metadata; and provide the technology to access
records of different Digital Libraries or systems,
ignoring the operating systems and details of
implementation of the information systems that
support them.
From a Computer Science point of view, resources
can be aggregated either with static or dynamic
methodologies. We consider static aggregation
protocols as methodologies that gather content from
405
Almeida P., Fernandes M., Arnaldo Martins J. and Sousa Pinto J. (2008).
RESOURCE AGGREGATION IN DIGITAL LIBRARIES - Static vs Dynamic Protocols.
In Proceedings of the Tenth International Conference on Enterprise Information Systems - DISI, pages 405-412
DOI: 10.5220/0001675604050412
Copyright
c
 SciTePress

different Digital Libraries or information
repositories, format it into a single and uniform
metadata language, and provide search facilities over
collected metadata/information. On the other hand,
we consider dynamic aggregation protocols as
methodologies that gather content from different
Digital Libraries or information repositories in real-
time. In the last methodology, the collected
metadata/information is obtained at the moment the
query is submitted to the search interfaces provided
by the systems that contain the information.
It is important to refer the role of service providers
and data providers according to each of these
methodologies. In the static model approach, the
data provider's only responsibility is to provide
metadata. The service provider must collect the
metadata, index it and provide search facilities.
In the dynamic model approach, data providers must
implement extra services, namely search facilities
over the metadata they contain. The role of the
service provider is less demanding, as, according to
this model, it does not need to store the metadata, or
index and provide search facilities over the indexed
content. Service providers act as pure information
aggregators.
Each of these aggregation methods has advantages
and drawbacks, but these concepts will be discussed
in more detail further ahead. At this stage, we intend
to clarify some of the concepts associated with each
methodology and, in the next section, we will
present some related work.
3 RELATED WORK
During the last years, several researchers from all
over the world have been studying the problematic
of aggregating information from different
repositories. The problems that arise are related with
the usage of different metadata to describe multi-
format digital object content and the different
technologies used to implement the information
systems that store and retrieve metadata from
repositories.
3.1 Z39.50
One of the first available protocols to search
simultaneously different databases is the Z39.50, an
American national standard for information
retrieval. It is formally known as ANSI NSO
Z39.50-1995 – Information Retrieval (Z39.50):
Application Service Definition and Protocol
Specification (National Information Standards
Organization, 2003). The main purpose of this
standard is to define a communication protocol to
access and query databases stored in different
computers with different software, facilitating the
process of interconnecting computer systems.
The standard specifies the formats and procedures
involved in the exchange of messages between a
client and server, enabling a Z39.50 client to request
the server to search a database, identify records
which meet specified criteria, and to retrieve some
or all of the identified records.
Z39.50 protocol also defines different record
syntaxes (Library of Congress, 2005), being most of
them variants of MARC records (Library of
Congress, 2007a) and resource format types, such as
mime-types and other file formats.
The Z39.50 Information retrieval protocol is
composed by a group of facilities to access database
information, namely: Initialization, Search,
Retrieval, Result-set-Delete, Access Control,
Accounting / Resource Control, Sort, Browse,
Explain and Termination. These operations specify
the interaction between the Z39.50 client and the
Z39.50 server, defining the services and functions
that can be invoked by the client.
In order to obtain the facilities that a Z39-50 server
supports, the Z39.50 clients can invoke the Explain
facility, and the server will answer with details of
the implementation, a list of databases available for
searching and the schema, record syntax and
element specification definitions supported for
record content retrieval.
Another operation defined in the Z39.50 is the
Browse facility. It is composed by a single service,
Scan. The Scan service is used to scan database
content, as long as the client provides an ordered
term list to scan (subject, names, titles, etc.), a
starting term and a number of entries to be returned.
Using these two facilities, Ray R. Larson developed
a Cross-Domain Information Server, using Z39.50
as the protocol to implement Distributed Resource
Discovery (Larson, 2001). As stated in the article,
the author used the Z39.50 functionality Explain
Database to determine the databases and indexes of
a given server. Then, using the SCAN facility, the
author extracted the contents of the indexes and used
that information to build “collection documents”.
The records were retrieved using probabilistic
retrieval algorithms. Z39.50 also defines a network
protocol to transfer information between the client
and server. Usually, the port number 210 is defined
as the default port for Z39.50 message transfers.
Using port 210 in modern Information Systems
causes a problem in large networks. As network
ICEIS 2008 - International Conference on Enterprise Information Systems
406

security is always an important factor to guaranty in
corporation networks, accessing a Z39.50 server on
the Internet imposes a reconfiguration of the firewall
in order to open port 210 to communications.
Sometimes this might be hard to negotiate with the
network administrator.
Other drawback associated to the network protocol
defined in Z39.50 is the fact that the messages are
transferred using rigid syntaxes. XML would
facilitate the interpretation of these messages by
third party tools.
On its essence, Z39.50 is a dynamic aggregation
system, since using this protocol Z39.50 clients can
search different databases and integrate the results
from distinct data sources. But in the case of the
system develop by Ray R. Larson, Z39.50 is used
just to collect metadata and build a database with
information from different providers. This way, we
consider that this system is a static aggregation
system, since in order to retrieve information from
different repositories it is necessary to first collect
data from the associated databases.
3.2 OAI-PMH
Another research group that studies the issue of
digital library metadata aggregation, created one
important initiative that has been gathering a
significant amount of supporters from the Library
and Archive community: the Open Archives
Initiative (OAI) (Open Archives Initiative, 2007).
This research group has defined a protocol to gather
information from different data sources, the Open
Archive Initiative Protocol for Metadata Harvesting
(OAI-PMH) (Open Archives Initiative, 2002).
This protocol specifies a standard mode to harvest
and retrieve records from repositories that
implement OAI-PMH.
OAI-PMH specifies two types of participants: the
data providers and the service providers. Data
providers are digital repositories that expose the
metadata about their objects with specific methods
defined in the OAI-PMH protocol. Metadata is then
gathered by harvesters, or aggregators. In the OAI-
PMH definition, these agents that gather the
information are defined as service providers.
Some important difference between OAI-PMH and
Z39.50 are the network protocol used for message
transfer and the syntax used for message transfer:
OAI-PMH uses HTTP, port 80, as a network
protocol, and uses XML to transfer messages
between the client and server. As port 80 is used by
Web Browser to access the Internet, usually this port
is always configured in the firewalls as an open port,
eliminating the need to open special ports in the
firewall.
By using XML, OAI-PMH messages can be easily
interpreted and processed by third party tools.
Data providers expose their metadata as sets: a
collection of metadata available for harvesting. They
are responsible to maintain a service available,
usually as a web server, which supports the OAI-
PMH protocol as a means of exposing metadata
from repositories. At least, data providers must be
able to expose metadata expressed in the Dublin
Core (DCMI, 2007) format, but more complex
metadata may also be disseminated.
The Dublin Core metadata is composed by fifteen
elements which are: contributor, coverage, creator,
date, description, format, identifier, language,
publisher, relation, rights, source, subject, title and
type. Not all the elements are mandatory in the
metadata that is sent to the service provider, but it is
important that data providers expose all the metadata
that is available for each record.
Service providers gather metadata from a group of
data providers and provide search mechanisms over
the information gathered. To collect data from
different repositories, the service providers issue
OAI-PMH requests to data providers.
Data providers answer to these requests with the
metadata collection they contain. Service providers
use the metadata to fill an inter-repository database
that stores information from all the data providers
that are associated with him. This collection of
metadata from different repositories is used as a
basis for building value-added services, namely
providing search facilities over large collections of
metadata from different repositories.
An example of a service provider is OAIster
(University of Michigan, 2007), a search engine
developed by the University of Michigan Digital
Library Production Service.
Their goal is to create a collection of freely
available, academically-oriented digital resources
that are easily searchable by anyone.
OAI-PMH is also an example of a static aggregation
system for Digital Libraries. In order to search the
content of several repositories simultaneously,
service providers must first harvest the repositories
content. In small repositories this could be a simple
task, but when we consider huge repositories with
millions of records, harvesting such a database is
necessarily a time consuming task.
RESOURCE AGGREGATION IN DIGITAL LIBRARIES - Static vs Dynamic Protocols
407

3.3 SRU
Library of Congress serves as the maintenance
agency for one of the new standards that perform
search queries on Internet Databases:
Search/Retrieve via URL (SRU) (Library of
Congress, 2007b).
SRU acts as a search protocol for Internet search
queries, and uses a Common Query Language
(CQL) (Library of Congress, 2007c) to define query
syntax. It is composed by a SRU Request, and a
SRU Response. A SRU Request is a URI as
described in RFC 3986 (Berners-Lee, 2005). A SRU
is composed by a base URL and a search part,
separated by a question mark ("?"). One example of
a SRU Request would be the following URL:
http://z3950.loc.gov:7090/voyager?version=1.1&ope
ration=searchRetrieve&query=dinosaur. As it can be
easily understood, the expected result of this query
would be resources that contain the term "dinosaur"
in the associated metadata.
All records retrieved by an SRU Request are
transferred in XML. It is not necessary that the
records themselves are stored in XML, but they must
be transformed to XML before the transfer from the
server to the client. Another important characteristic
of SRU is that records in the response may be
expressed as a single string, or as embedded XML.
This protocol an example of dynamic aggregation
since the search results are obtained on real-time. As
we will show further ahead, the SRU protocol is
similar to the methodology that we have adopted and
that we will present in this chapter. The problem
with SRU is mainly associated with the limitations
imposed by specification, and the lack of a
mechanism to retrieve the records themselves. Using
SRU it is only possible to retrieve metadata
associated to a record, not the information that was
classified and indexed.
4 STATIC AGGREGATION VS.
DYNAMIC AGGREGATION
In this section we pretend to identify the
implications in using either static or dynamic
aggregation. Both of the methodologies have
advantages and drawbacks, and our intention is to
provide a list of features that will be easy (or
difficult) to achieve depending on the adopted
methodology.
Drawbacks of static aggregation:
1. Time to harvest the entire repository
content.
2. Searches are made in possibly outdated
metadata.
3. Difficult integration with other software
tools.
4. Harvested data must be stored in the server
that provides the search mechanisms.
Advantages of static integration:
1. Offline repositories can be queried.
2. Possibility to detect and eliminate
duplicate records in different repositories.
3. Search performance depends only of the
server.
Searching simultaneously several repositories from a
static set of information can originate some
problems. Retrieving the right information from
large collections of metadata is not an easy task. For
example, in this article (Hochstenbach, 2003)
authors defined some future guidelines for OAIster,
in order to improve the quality of information
retrieval. These guidelines include: searching within
institutions; browsing capability; eliminate duplicate
records (records that are the same among
repositories). These limitations belong to a specific
system, and this is the reason why they were not
considered in the list of drawbacks.
The time required to harvest data from a repository
might also be a problem. It is common to think in a
Digital Library as a repository that stores millions of
records. Harvesting all these records might be a time
consuming task.
In order to obtain updated results it is necessary to
collect new data from repositories in a regular basis.
If dealing with three or four distinct repositories is a
perfect acceptable number of repositories to keep
track of, considering the extensibility of these
systems, and the possibility to interact with a higher
number of information repositories, this type of
approach might lead to another time consuming task.
Regarding to the integration of harvesters by other
information systems we could not identify other
advantages in addition to reuse harvested
information. Since the information has to be
harvested and stored in a centralized mode, querying
directly data providers should be a better option than
integrate the search results in another software tool.
Another problem that does not encourage the
integration of harvesters is the lack of control in the
ICEIS 2008 - International Conference on Enterprise Information Systems
408

harvesting process by the system integrator. It is
difficult to determine if the results are updated or
synchronized with the original repository.
Yet another issue in static aggregation is the need to
store all the metadata collected in the harvester.
Since the data is already present in the original
repository, harvesting the metadata represents
creating redundant copies of the same information,
and as a consequence, the harvester will necessarily
need more resources, namely disk space, to store
metadata copies.
Despite the drawbacks already identified, static
aggregation has some advantages over real-time
dynamic aggregation. One advantage is that online
repositories can be queried, even if the service
providers, harvester, has no connectivity with the
data provider. Since all the metadata is stored in the
server, queries may retrieve result sets from o²ine
repositories. Yet, it is necessary to be aware that this
might lead to another problem: the harvester might
provide the link but the document might be
inaccessible.
Identifying duplicate records is, most of the times, a
time consuming task. Since a harvester usually takes
some time to gather information from a repository, it
is acceptable to spend small amounts of time to find
duplicate records. As a benefit, a user could get
results from different repositories ignoring duplicate
results.
Regarding to the drawbacks and advantages of
dynamic integration we have identified the
following items:
Drawbacks of dynamic aggregation:
1. Offline repositories won't be able to show
results.
2. Difficult to eliminate duplicate records.
Advantages of dynamic integration:
1. Guaranty of updated results.
2. Direct integration in software tools.
3. Web Services and XML oriented,
providing high interoperability and low time
cost in application development.
The drawbacks associated with dynamic aggregation
are related with connectivity and processing time
issues. The first problem in dynamic aggregation is
that if a metadata provider has no connectivity with
the system that queries the repository, the aggregator
system will not be able to obtain search results. The
second problem is related with eliminating duplicate
records. In order to detect duplicate records it is
necessary to spend some processing time. This
might constitute a problem because it would mean
spend extra time to obtain the results. In our opinion,
finding duplicate records in an aggregator system is
not a critical problem, but despite our point of view,
it still is a limitation in real-time aggregation
systems.
One important advantage is the guaranty that the
obtained results are always updated. As in dynamic
aggregation data providers implement search
interfaces, service providers always have access to
the most actual information. This is an important
issue in repositories that have high rates of
document insertion, like the case of the repository of
the University of Aveiro, where every day new items
are being added or updated in the different
collections.
Regarding to advantages 2 and 3, they are only
possible if the dynamic aggregation system is
implemented using Web Services and XML. These
two technologies provide important an important
feature to the information systems: interoperability.
It also allows different computers, with different
operating systems and software, to exchange data
via a common set of business procedures.
Integrating a search interface of a data provider that
exposes the search functions with Web Services is a
very rapid, simple and efficient mode to put two (or
more) applications communicating with each other.
5 CASE STUDY
The case study that we present is based on the
information systems strategy adopted by the
University of Aveiro. The policy being implemented
inside the institution promotes the development of
software and information systems that expose data,
metadata and search interfaces via Web Services.
The actual philosophy is to maintain the information
centralized in specialized service providers, but
accessible to the rest of the information systems of
the university, avoiding data replication between the
different databases that exist over the University.
Three systems compose the actual Digital Library:
SInBAD, Curriculum and eABC. Each of the
systems contains specialized information that, by
itself, is very useful for the university. But the
increased value obtained with the integration of
these three systems is one interesting example of the
potential of growth of functionalities and services
just by using dynamic aggregation.
RESOURCE AGGREGATION IN DIGITAL LIBRARIES - Static vs Dynamic Protocols
409

5.1 SInBAD
SInBAD (Almeida, 2006) is an integrated system for
digital libraries and archives. This project is being
developed at the University of Aveiro and its major
goal is to collect and store information about
institutional records.
At the moment, this system contains a collection of
posters, a photographic archive, an audiovisual
archive, a thesis collection and a vast collection of
bibliographic information of jazz books, magazines
and records.
One interesting feature of SInBAD is that it contains
four subsystems, which can be accessed individually
or simultaneously. It implements a real time
dynamic aggregation mechanism internally,
providing a search portal that can simultaneously
access different types of information.
5.2 eABC: Bibliographic Archive for
Scientific Production
eABC (Santos, 2005) is an information system that
contains bibliographic records of the scientific
publications of the University of Aveiro. Initially,
the main purpose of the system was to provide a tool
where researchers could manage their scientific
publications. Later, the Research Institute began to
use this system to manage the research production of
the University of Aveiro.
Every year, the Research Institute produces a report
with all the publications of the research units that
belong to the University and sends it to the
Portuguese Foundation of Science and Technology.
This system provided an easy mode to accomplish
the difficult task of gathering publication
information of all the researchers inside the
university.
5.3 Curriculum
Curriculum (Teixeira, 2005) is an information
system that stores the curriculum vitae (CV) of
researchers, professors, students and other elements
of the University of Aveiro. This system has one
important objective as an element of the University
of Aveiro information sources: maintain an updated
repository with the researcher's CVs. Despite the
simple goal of this system, the quality and coverage
of the services that it can provide make of it one of
the key information systems inside the university.
Over the last two years, the candidates of the PhD
scholarship applications of the University used
Curriculum to store their CV. Another example of
system usage is that when researchers of the
University need their CV to be attached to a project
or career progress contest they can access this
system and obtain an updated version of their CV.
5.4 Real-time Resources Aggregation
In order to build increased value services
implemented with the existing systems, we have
defined an interaction schema between SInBAD,
eABC and Curriculum. Each one can perform its
tasks without interacting with other external
systems, but the interaction between the three
systems allows users to access transparently a vast
group of related information.
Figure 1: Dynamic aggregation model.
Figure 1 depicts the interaction diagram between the
three systems.
SInBAD provides digital content objects, eABC
provides bibliographic information and Curriculum
provides author information. The aggregation of
resources is possible due to the Web Services
interfaces that all systems implement. Each of these
interfaces has specific methods to insert, delete,
search and retrieve data and metadata from any
system.
As it becomes obvious, the increased value services
come from the real-time dynamic integration of the
data and metadata from the three information
systems.
Consider, for example, a situation where a
researcher inserts a new publication in eABC.
Before the existence of SInBAD, the user could not
upload a PDF file of his scientific work. The only
option that was available was a bibliographic field
where the user could register the URL of a digital
version of the article. Now, eABC allows the user to
select a PDF file with the work content and submit it
ICEIS 2008 - International Conference on Enterprise Information Systems
410

along with the bibliographic information. Actually,
the PDF file is not stored in eABC, but in SInBAD.
Each of the systems maintains its primary goal, and
there is no overlapping of functionalities. As an
increased value, all the scientific publication will be
stored in a centralized repository, avoiding the
dispersion of publications for several user machines.
Another interesting point is that, when SInBAD is
queried for scientific publications, the search results
are obtained from eABC. SInBAD just identifies
articles that are stored in its database and presents
the users a list of scientific publication that meet the
required query parameters. As the query is actually
resolved by eABC, it is guaranteed that the most
actual metadata is always accessed. This eliminates
the time consuming task of synchronizing metadata
values from different information systems.
Based on the author ID stored in eABC, it is also
possible to link to the CV of the researchers that
produced the scientific publication. Once again, the
complete CV is stored in two different systems,
because the scientific publications associated to the
researcher are stored in eABC. The Curriculum
system just stores personal and professional
information. This characteristic is very important in
large research units. Sometimes the number of
authors associated to a publication is considerably
large, and easily one or two authors could forget to
associate the scientific publication with his CV.
Therefore, once the record is created in eABC, the
association with the authors CV is automatic, which
alleviates researchers from the tedious task of
updating their CVs.
As all the services are interconnected and related,
users accessing one of this systems can easily access
a vast group of data that is related. It is also possible
to build new services that dynamically import data
from the three systems and present subgroups of
information. As the systems began to operate,
several web masters responsible for the web pages
of the departments of the university began to request
us information about the scientific publications
associated to the department units, links to the
researchers CVs, etc. For example, some
departments in the university wanted their web
pages to contain a list of the master and PhD thesis
and scientific articles related with the department.
Before the development of SInBAD and eABC, this
list was written inside the department web page, or
was retrieved from a small database. Now a
complete list with this information is retrieved
directly from SInBAD an eABC, allowing the web
masters to integrate updated data in the department
web pages.
6 CONCLUSIONS
We consider that the new XML dynamic
aggregation based technologies that allow different
systems to communicate in an integrated format
might take an important role in metadata and
resources aggregation. Web Services provide all the
necessary means to build interoperable repositories
that share interfaces to their search mechanisms.
Instead of building system providers, according to
the OAI-PMH definition, that must deal with the
harvesting, search and aggregation of the
information, a new generation of system providers
might only deal with the aggregation of the results.
Data providers assume higher responsibility in this
new proposal, being responsible not only to manage
the metadata, but also to provide query interfaces to
system providers.
The new strategy for the information systems
architecture of the University of Aveiro Digital
Library is actually giving its first results. As the
aggregation of different content from different
repositories is becoming a rapid and warrantable
process, all the web developers inside the university
are exploring the new possibilities of accessing
department specific information.
Resource aggregation is obviously an important step
to increase the potential of Digital Libraries and
Institutional repositories.
REFERENCES
National Information Standards Organization, 2003.
ANSI/NISO z39.50-2003: Information retrieval
(z39.50): Application service definition and protocol
specification, NISO Press.
Library of Congress, Network Development & MARC
Standards Office, 2005. Registry of Z39.50 object
identifiers. Last update: October 2005, Available at
www.loc.gov/z3950/agency/defns/oids.html. Accessed
in November 2007
Library of Congress, Network Development & MARC
Standards Office, 2007a. Marc standards. Last update:
July 2007, Available at http://www.loc.gov/marc/.
Accessed in November 2007
Larson, R. Ray, 2001, Distributed resource discovery:
Using Z39.50 to build cross-domain information
servers, In JCDL 01. ACM Press.
Open Archives Initiative, 2007. Available at
http://www.openarchives.org, Accessed in November
2007
Open Archives Initiative, 2002. The Open Archives
Initiative Protocol for Metadata Harvesting, Last
update: October 2004, Available at
RESOURCE AGGREGATION IN DIGITAL LIBRARIES - Static vs Dynamic Protocols
411

http://www.openarchives.org/OAI/openarchivesprotoc
ol.html, Accessed in November 2007
DCMI, 2007. The Dublin Core Metadata Initiative, Last
update: November 2007, Available at
http://dublincore.org/, Accessed in November 2007
University of Michigan, 2007. OAIster, Available at
http://www.oaister.org/, Accessed in November 2007
Library of Congress, 2007b. SRU: Search/Retrieve via
Url, Last update: September 2007, Available at
http://www.loc.gov/standards/sru/, Accessed in
November 2007
Library of Congress, 2007c. CQL: Contextual Query
Language (SRU Version 1.2 Specifications), Last
update: August 2007, Available at
http://www.loc.gov/standards/sru/specs/cql.html,
Accessed in November 2007
Berners-Lee, T., 2005. Uniform Resource Identifier (URI):
Generic syntax, http://www.ietf.org/rfc/rfc3986.txt.
Hochstenbach, P., Jerez, H., Sompel, H., 2003. The OAI-
PMH static repository and static repository gateway,
In Proceedings of the Joint Conference on Digital
Libraries (JCDL03), IEEE Press.
Almeida, P., Fernandes, M., Alho, M., Martins, J., Pinto,
J., 2006. SInBAD - A digital library to aggregate
multimedia documents, In AICT-ICIW '06, IEEE
Press.
Santos, J., Teixeira, C., Pinto, J., 2005. eABC - Um
Repositório Institucional Virtual, In XATA 2005 :
XML Applications and Associated Technologies, pp.
40-51, University of Minho Press.
Teixeira, C., Pinto, J., Santos, J., 2005. Curriculum@ua -
Online XML based personal curriculum, In XATA
2005 : XML Applications and Associated
Technologies, University of Minho Press.
ICEIS 2008 - International Conference on Enterprise Information Systems
412