page 1 of 12
Developing further the concept of ePortfolio with the use of
Semantic Web Technologies
Veronika Hornung-Prähauser, Wernher Behrendt, Motti Benari
Salzburg Research Forschungsgesellschaft
Jakob Haringerstrasse 5/3
This paper is in a series prepared for the Special Session
“ePortfolios: Technical challenges and new developments”
Paper I “ePortfolio for a Learning Society” - ePortfolio concept, trends and challenges
Paper II “ePortfolios at the heart of eLearning interoperability” – Issue(s) of organisational
and technical interoperability
Paper III: “Developing further the concept of ePortfolio with the use of Semantic Web
Technologies”: Ontology based ePortfolio templates for semantic interoperability
Key words: ePortfolio, Semantic Web Technologies,Ontologies, Knowledge
The paper deals with the issue of further developing the concept of ePortfolios in terms of
semantic interoperability. Social Software and Semantic Web Technologies may enrich the
traditional concepts of ePortfolios. The purpose of this contribution is to summarize current
research issues and to describe basic steps needed for a knowledge based management of e-
portfolios at a large scale. Specifically, it proposes the method of using an ontology driven
ePortfolio-template that would allow machines to more flexibly read and re-use the ePortfolio
contents. This may open up new application scenarios of ePortfolio-software for educational
institutions ( e.g. web-based processing of educational standards) and for enterprises (e.g.
online- recruitment methods).
ePortfolios already have a much longer tradition in the educational sector of anglo-american
countries than in mainland Europe, except UK  . However, lately both the paper based
portfolio concept and its digital version are seen as important tool supporting the learning,
assessment and presentation process of competences and skills. In technical terms, the digital
e-portfolio can be seen as an information object that uses electronic media (docs/pdf;
pictures/jpg, videos, audio-files, html) and services. ePortfolios need to be managed by either
a dedicated ePortfolio management system or by an extension to a content management
system (CMS). One can distinguish three different layers (Baker, 2005, p.4):
? The repository – all the artefacts (or links to) and reflections of the eP owner that will be
used to produce presentations
page 2 of 12
? The presentations – all the views that are accessible, publicly or privately
? The services – all the features useful to provide support to learning, assessment etc.
(webblog, RSS Feeds)
The learner builds and maintains his/her digital repository of different files and uses it to
demonstrate competence and reflection on their learning in different domains. Having access
to their records, digital repository, feedback and reflection students can achieve a greater
understanding of their individual growth, learning and career planning. It is regarded as a
powerful tool for accreditation for prior and/or extra-curricular experiences and control over
access (see also Baker, 2005 ). The interesting issue at stake here is that while in former
years ePortfolio were produced only as simple websites, the fast changing technologies for
semantic web publishing will offer new methods and ways of creating, managing and
exploiting the collection of “Portfolio artefacts”.
In this paper the model of Knowledge Content Objects (KCO)  shall be introduced. We
will discuss whether the KCO could be a contribution to semantically enriched e-portfolio
collections. The model was developed in two recent EU projects by the research group of
Knowledge Information Systems at Salzburg Research, Austria , together with partners in
those projects (CULTOS - Cultural Units of Learning - Tools and Services and METOKIS
Methodology and Tools Infrastructure for the Creation of Knowledge Units). The KCO can
be regarded as a domain independent ordering mechanism to combine individual content
resources via meaningful relationships expressed in a knowledge model. The initial
applications were developed for educational settings and the KCO concept is at the moment
applied to different business domains. Based on our recent research we suggest that the model
may also be applicable in the world of digital portfolios and has the potential to enable more
"intelligent" use of e-portfolio software. As is the case with learning objects, the question of
re-use  needs to be asked also with respect to e-portfolios:
Can a query based on a pre-defined ePortfolio ontology (via a template)- help to
better search and re-use (public parts) of a huge number of ePortfolios? This may be
of use for example for the management of an educational institution (e.g. for
recommendation of an internship; for alumni networks)?
Is it possible to automatically match different ePortfolio types with open job
advertisement of companies, public labour offices or educational institutions
according to a pre-defined ontology and learners`profile data? How can the aspect of
“living portfolios” be taken into account?
2 The value of semantic web technologies for enriching the
method and use of ePortfolios
2.1 Understanding the research challenges of ePortfolios and semantic web
Darren Cambridge, one of the most well-known visionaries on ePortfolio technologies
proposes that with the emergence of the semantic web, a better framework for sharing and
re-using data on the web is given and only then, can ePortfolios serve as a tool for connecting
people based on interests, skills, or other criteria if presented in the right way . He points
out five major technical challenges for electronic portfolio technologies: Design, Semantics,
Factoring, Community and Decentralization. Given the purpose of this paper we concentrate
only on the challenges related to “Semantic”.
page 3 of 12
What is meant by the “Semantic Web”?
In general the Semantic Web is regarded as a project that intends to create a universal
medium for information exchange by giving meaning (semantics), in a manner
understandable by machines, to the content of documents on the Web. Currently under the
direction of the Web's creator, Tim Berners-Lee of the World Wide Web Consortium, the
Semantic Web extends the ability of the World Wide Web through the use of standards,
markup languages and related processing tools Currently, the World Wide Web is based
primarily on documents written in HTML, a language that is useful for describing, with an
emphasis on visual presentation, a body of structured text interspersed with multimedia
objects such as images and interactive forms. However HTML has limited ability to classify
the blocks of text on a page, apart from the roles they play in a typical document's
organization and in the desired visual layout. So with semantic web technology, it should
become possible to search for a specific type of e-portfolio, or for a specific set of
demonstrated skills. Such targeted search is at present only possible in specialised databases
or with bespoke systems specialised at e-portfolio management.
The Semantic Web addresses this shortcoming, using the descriptive technologies RDF and
OWL, and the data-centric, customizable markup language XML. These technologies are
combined in order to provide descriptions that supplement or replace the content of Web
documents. Thus, content may manifest as descriptive data stored in Web-accessible
databases, or as markup within documents (particularly, in XHTML interspersed with XML,
or, more often, purely in XML, with layout/rendering cues stored separately). The machine-
readable descriptions allow content managers to add meaning to the content, thereby
facilitating automated information gathering and research by computers.
Why is it necessary in the field of ePortfolio to aim for semantic interoperability?
One of Tim Berners-Lee’s initial visions for the Semantic Web (Tim Berners-Lee, James
Hendler, Ora Lassila, The Semantic Web, Scientific American, May 2001. ) is the desired
ability for a personal digital assistant to automatically query a general practitioner’s office and
create an appointment time that works for all parties. Mapped on to the domain of ePortfolios,
this could mean that for example a search in the alumni database will automatically match
with open job adds in online job databases.
In another scenario of formal education, a college system tells the student’s tool what needs
to be included in her application portfolio, and the student’s tool knows how to use this
information to facilitate the selection and organization of materials. The student’s tool adapts
its interface to scaffold the process of reflecting on learning based on guides it requests from
the college system. The student’s tool publishes a portfolio with artifacts and reflections
mapped to competencies in a format the college system can analyze, allowing it to compare
the student’s learning with that of hundreds or thousands of other students, providing
aggregate data for use in programmatic and institutional assessment. In addition, the college’s
portfolio community server knows how to search through alumni portfolios for evidence of
team-building ability useful to our student. Future e-Portfolio tools need to represent portfolio
knowledge that both people and computers can understand (Cambridge, 2003 pp.10-16).
Imagine a student graduating at the age of 25 from university. She has registered her
portfolio at the server of the university and then moves on to her first job. Then she would
like to add the learning experience of her intern-ship at a company and her profile of the
alumni association. In the course of a professional learning career one will want to have more
than one “view” on one's profile: this means using the portfolio data for different purposes
and fill in different portfolio types with the same content. One would like to flexibly combine
page 4 of 12
parts of one personal portfolio-view with an institutional one (e.g. education authority) and
moreover one would like to have a guarantee that updates/versions are recognised
At present, the general purpose Web design tools most commonly used to construct Portfolios
(for a list of tools see http://www.eife-l.org/portfolio/inventory/English 10092005), afford
authors a great deal of control over what a human experiences when reading their portfolio
(Barrett 2003). While it requires a significant investment of time and a relatively high level of
technical and visual skill, using tools like Photoshop, Premiere, Acrobat, Dreamweaver, and
Flash, learners can create visually and textually sophisticated representations of their learning.
However, to our knowledge none of these tools provides a mechanism for “knowledge
representation”, i.e. for expressing meaning through structures that computers can parse and
produce through reasoning over eportfolios.
Figure 1: Role of ePortfolio in a Learning Management System; Graphics adapted from Dessi N., 2004.
SRS / MIS
record / manage
HTML, XHTML, XML
SVG, SMIL, JPEG,MPEG
ODRL METS ZThes
2.2 Using ontologies for structuring ePortfolio collections and -processes
Tom Gruber of Stanford University has the shortest and often-cited definition: "An ontology
is a specification of a conceptualization.". Gruber then elaborates the definition as follows:
"The term is borrowed from philosophy, where an Ontology is a systematic account of Existence. For AI
systems, what "exists" is that which can be represented. When the knowledge of a domain is represented in a
declarative formalism, the set of objects that can be represented is called the universe of discourse. This set of
page 5 of 12
objects, and the describable relationships among them, are reflected in the representational vocabulary with
which a knowledge-based program represents knowledge"1
In other words, these formal ontologies, defined as data structures of a specific knowledge
community/domain2, can serve as “guidance” for machines through the rich world of personal
competences, skills and accrediations.
The true value of ontologies when used for ePortfolios may lie in binding together relevant
parts between content and an ePortfolio ontology so that raw content is enriched with more
formal “meanings” pre-defined in a shared ontology. Enriching content is also termed as
knowledge binding which depends upon human effort to tag the content, thus making the
knowledge accessible. Key mark-up languages, such as XML, RDF and OWL are often
chosen to represent the semantics via ontologies and make them machine accessible and
How to create an ontology - what are the inputs for an ontology based ePortfolio template?
Given our definition from above for a formal ontology, we need to relate the concepts which
are commonly known and used by humans when talking about e-portfolios, to a formal
representation of these terms. Such a representation is given to us by a so-called upper
ontology (some prefer the expression of "foundational" ontology which indicates the need for
basic representational primitives such as time, space, events, actors and roles). This process is
called ontology alignment: our private ontology is made compatible with a more
comprehensive or more basic ontology. Once this is done we can in principle, start adding
content to our database and we can semantically annotate that content by using the ontology
terms defined. However, this would have a price: we now have to build an application
"around" our specific ontology, so that information items can be retrieved much in the same
way as one queries a relational database. With the KCO we have provided a generic content
and knowledge item, for which many basic mechanisms are already pre-defined and therefore,
we suggest that taking the extra modelling effort is probably worth doing.
2.2.1 A taxonomy of e-portfolios
The EPPIC research group on ePortfolios categorize the different types of eportfolios
according to its didactical purposes6. Four common ones are distinguished below.
• Assessment portfolios: Portfolios used for assessment purposes only are usually
organized around items such as the candidates’ products, evaluations, photographs
• Showcase portfolios: When persons compiling a portfolio are free to determine the
content of their portfolios, they most often tend to display examples of their best
work or evaluations of that work. Such portfolios are usually referred to as
showcase portfolios and resemble those compiled by artists and architects (e.g.
1 Tom Gruber, http://www-ksl.stanford.edu/kst/what-is-an-ontology.html (last visited, September 2005)
2 In information science, an ontology is the product of an attempt to formulate an exhaustive and rigorous
conceptual schema about a domain. An ontology is typically a hierarchical data structure containing all the
relevant entities and their relationships and rules within that domain (e.g., a domain ontology). The computer
science usage of the term ontology is derived from the much older usage of the term ontology in philosophy.
page 6 of 12
• Development portfolios: A portfolio may also be designed as an instrument to
keep track of and plan the owner’s development. In such a case it is referred to as a
development portfolio. The point of departure here could be a summary of what
the owner should master in order to be certified.
• Reflective portfolios: When portfolios are used for the purposes of monitoring the
owners’ development, it is important to know how he evaluates and analyses
himself. Therefore it is crucial that portfolios used in this way contain written
reflections by the owner.
• Combinations: Portfolios are usually used for a combination of different purposes
and combine characteristics of each of the typical portfolios described above.
2.2.2 Using an ontology to define the structure of ePortfolio content
The structure of an ePortfolio can now be defined in an ontology. We use the overview from
Cambridge B.  for describing what work will be needed to develop an ePortfolio ontology.
• Categories – used as building parts of an ontology description
Portfolios are typically separated into sections, and thus it follows that ePortfolios should as
well. These categories can also have sub-categories and sub-sub-categories ad infinitum.
Since we cannot intelligently predict the deepest tree structure used in a portfolio, thus the
ePortfolio system should not constrain it. User groups have very different requirements for
what categories or topics their ePortfolios need to cover. Some are highly specified others are
very open. To accommodate both solutions the system has a way of allowing preset categories
while not requiring the use of the presets so all the user groups are satisfied.
• Fields - represent Data and Metadata
Within the portfolios each portfolio category contains a reflection, summary or justification.
Sometimes they contain all three, sometimes they contain a description of the artifact and
sometimes they are simply a collection of artifacts. Anything written outside of an artifact is
regarded as field. A field may be directly related to an artifact, it may be simply an
introduction to the category, or it may not be there at all. In short, not enough is known these
categories, thus an ePortfolio system should provide flexibility. Similar to categories, each
user group has different needs for fields in the portfolio.
• Artifacts – represent Media files
Integral to ePortfolios are artifacts; proof of the author’s work, skills, interest, or knowledge
the portfolio may be focusing on. An artifact in an ePortfolio system can be anything digital.
Preferably, an ePortfolio system will be able to handle any type of artifact from a simple URL
to an executable file (video, audio, picture etc., weblog) created by the author.
• Portfolio Profiles: can serve as Metadata-data = Ontological description
Portfolio profiles are more complex. While the content of the portfolio takes the form of
categories, fields, and artifacts, the portfolio profile covers what type of portfolio the author is
creating. If all or some of the portfolio content is to be produced as a complete professional
and public portfolio, then the particulars of that can be defined and classified as part of the
portfolio profile. Users want to be able to create multiple types of portfolios while not having
to have separate accounts for each.
Whereas the above items are needed to structure the ontology itself, there are further issues to
be addressed when it comes to possible usages of such portfolios. The portfolio software
page 7 of 12
should provide functionalities such as supporting a discourse, the option to give/allow
feedback, or to use announcement functions to keep portfolios up-to-date even if they are
spread over several databases.
2.2.3 Example of an ePortfolio low level and upper level ontology
In the project SPARCS (2004) , the software developers were trying to build such an
eportfolio ontology which was firstly intended to serve the British Columbia University
(Canada) and then intended to be used as generic one for all different domains using
ePortfolios (upper-level ontology). They created the following upperlevel-ontology, shown in
the graph below. (only the classes and their relationships are represented in the figure. Slots;
also known as fields or parameters are omitted for clarity). As the developers point out in
order to be useful an upper level ontology must support extensions for particular purposes .
The upper level ontology offers starting points for all of the lower level concepts. Assessment
and grading have a specialization relationship with Feedback. Different organization schemes
can be built off the portfolio and categories. The authors evaluated the success of the upper-
level ontology (ULO) by extending it for each of the domains from which we generated it.
We were successful in building back from the ULO to each of the specific domains, though it
requires more effort for those using unusual terminology such as the BC Ministry guidelines.
Upper level Ontology for ePortfolio as proposed by SPARC
Figure 2: Brokenshire, D., Bogyo, B. and Kumar, V Simon (2004, p.4).
What can be gleaned from the SPARCS ontology is that its upper level would appear to be
more a pragmatic choice than a strongly founded upper level ontology as understood by
practitioners in the Semantic Web community. The full SPARCS ontology was not accessible
to us at the time of writing, as it was not yet on-line. In our view, SPARCS is a first attempt at
defining the fundamental categories needed to describe e-portfolios, but would benefit from
proper alignment with one of the better known upper level ontologies, be they Sowa's
ULO, Gangemi et al's DOLCE, or SUMO (Niles 2001). However, from our own
experience it is at present neither an easy task to align one's domain ontology with an upper
page 8 of 12
level ontology, nor is there at present, a clear cut "business case" for doing so. Nonetheless,
we would concur with Brokenshire et al that making the effort is likely to have beneficial
effects longer-term, and may even be a good investment medium-term, because the
ontological analysis often reveals "hidden semantics" that need to be clarified for use in
3 Semantic interoperability with Knowledge Content Objects
The following section describes how we intend to integrate a domain model for e-portfolios
into a carrier structure (the aforementioned KCO) and how the use of this carrier structure
can facilitate the definition, creation and management of ePortfolios. We also explain the
relationship between KCO, ePortfolio and upper level ontology using DOLCE as the
3.1 Definition of a KCO
A Knowledge Content Object is a conceptual schema for arranging and wrapping content for
the semantic web. It is defined using a so-called foundational ontology as its reference and it
enables dynamic gradual evolution from a generic schematic prototype into firstly, an object
which is aware of educational concepts (e.g. the notion of a media resource or a course
module) and which can be contextualized and enriched by further semantic annotations (i.e.
knowledge structures) in a seamless fashion.
A Knowledge Content Object comprises of six facets:
Figure 3: KCO Structure Westenthaler R. et. Al
Three are of particular interest to the educational sector:
1. The digital content (media resources) and propositional content expressing what the
content is about (at arbitrary levels of detail, not just by way of keywords)
2. A community description which indicates how the content object can be used in
principle and which is capable of keeping a record of how it has already been used.
page 9 of 12
3. Sets of (machine readable) indications on how and under what terms this data can be
accessed (these include license terms and access and usage rights).
Since each individual KCO can contain the aforementioned indications, any application that is
familiar 3 with the concept of KCOs, can identify, without any specific pre knowledge, the
KCO’s particular relevance and can decipher what manipulations in its own scope can be
done with the content.
These sets of indications are organized in facets and sub-facets. For instance, there is a
business description facet dedicated for business and legal semantics (license schemes, usage
negotiation protocols, pricing scheme, electronic contracts), a community description facet
dedicated for user-tasks-description, (what tasks can it be utilized for) user-community-
description (what type of users benefit from it) and user-history description (versioning data,
upgrading data, etc.). Other facets are designed for presentation description, and for trust and
security description (access semantics). The KCO structure is elaborated elsewhere by
Behrendt, Gangemi, Maass, & Westenthaler (2005).
3.2 KCO Base Design – DOLCE as foundational ontology
The core structure of the KCO is built around the semiotic ontology design pattern provided
by the DOLCE foundational Ontology, which is shown in the following graphic.
Figure 4: DOLCE Information-object design pattern
3 For an application to be familiar with KCOs we devised an open distributed architecture and a
communication protocol. The architecture constitutes a "middleware" that can be employed to make
existing applications "KCO aware".
page 10 of 12
Without going into detail, every KCO is a derivative of the "Information Object" in the
DOLCE Ontology, and therefore inherits all the semantic relationships that are defined for the
Information Object, as indicated in the diagram above.
Moving back to the KCO and its internal structure, the most interesting facet is the so-called
domain description in which the KCO holds two things: firstly, a reference to a specific
domain ontology - in our case a proposed semantic model of eportfolios - and secondly,
semantic descriptions which are related to the actual content of somebody's actual e-portfolio.
The KCO exists, to start with, as a template structure that is only primed by the e-portfolio
ontology. When we add actual content and annotate that content with reference to the priming
ontology, then we create a fully instantiated knowledge content object - in this case a full e-
portfolio with its content accessible through the semantic descriptions.
Figure 5: Information-Objects and Decriptions of the Content Description Facet [11, p.9]
As can be seen from the above diagram, the KCO content description facet gives us a scaffold
along which we can add content, segment the content arbitrarily (in descriptions) and then
annotate each of the segments or groups of segments as we wish.
3.3 KCO Carrier structure as solution for interoperability
We should indicate here at least the principles behind the structure of the propositional
The KCO carries a list of media references which are intended to point to real media files.
Thus, the collection of these referenced media files is actually, the full intended content of the
KCO. In order to make this content accessible for machines, the propositional content facet is
linked to a domain ontology, which represents the Universe of Discourse for all content
descriptions of this KCO. This enables to associate selections of the multimedia assets with
page 11 of 12
statements that are valid according to the ontology. The semantic annotation is very flexible
as it can relate to segments of a media asset (eg, a scene in a video or a region in an image) or
even to a relationship that holds between some media assets.
The KCO supports the development from an elementary generic structure to an instantiated
contextualized and personalized tradeable KCO as described in figure 2.
Figure 6: Levels of KCO abstractions and KCO facets (Behrendt et al 2005).
To summarise, we also believe that using semantic web technologies could be a boon to the e-
portfolio movement and vice versa, e-portfolio could prove an important stepping stone
towards widespread acceptance of the usefulness of semantic web technology.
4 Challenges ahead
Digital portfolios could be an excellent testing ground for some of the possibilities semantic
web technologies offer for solving problems in terms of semantic interoperability. However,
three challenges using KCOs lie ahead:
• Semantic annotations need agreement on a commonly defined and accepted
ePortfolio ontology: Developing an agreed basic ontology for ePortfolio
collections would increase the level of interoperability between people,
organisations and systems (e.g.LMS systems).
• Development of domain specific sub-ontologies for different types of portfolio
(e.g. Arts portfolios vs banking or professional development schemes)
• Guarantees for storing eportfolios “lifelong" because every citizens must have a
right to his/her educational identity. In this point, e-portfolio and many other
applications meet in mutual agreement, because reliance on the existence and
persistence of data is so fundamental to the Information Society that it should be
declared a basic citizens right.
page 12 of 12 Download full-text
 Layte M. : ePortfolios – State-of-the-Art of ePortfolio at European and International Level, Abstract
for Keynote at ePortfolio Forum Austria, 2005, also www.eportfolio.salzburgresearch.at .
 Baker, A: ePortfolio Tools and Services: State of the Art of ePortfolio technologies. Abstract for
Keynote at ePortfolio Forum Austria, 2005, also www.eportfolio.salzburgresearch.at .
 Motti B., Behrendt W., Hornung-Prähauser V., 2005 in: (currently under review for the Special
Issue on Advances of Semantic Web for E-learning: Expanding learning frontiers, British Journal
of Educational Technology; Edited by: Nick Rushby , Print ISSN: 0007-1013, Online ISSN: 1467-
 Tao F., Davis H., Millard D., Woukeu A, The Semantic Aspects of e-Learning: Using the
Knowledge Life Cycle to Manage Semantics for Grid and Service Oriented Systems, p.5, 2005,
School of Electronics and Computer Science University of Southampton Southampton, UK.
 Cambridge D, and Cambridge B. :The Future of Electronic Portfolio Technology: Supporting
What We Know about Learning Keynote address presented at ePortfolio 2003 (Scenarios): The
first international conference on the digital portfolio, October 10, 2003, Portier, France Cambridge
 Tartwijk v.J., Driessen, E. Exploring Different Types of Electronic Portfolios Jan, EPICC Team,
26. April 2004, http://www.epiccproject.info
 Brokenshire, D., Bogyo, B. and Kumar, V Simon Fraser University, Surrey, BC, Canada
Towards an Upper-Level Ontology for Information Exchange in ePortfolios
 John Sowa. Knowledge Representation: Logical, Philosophical, and Computational Foundations.
 A. Gangemi, N. Guarino, C. Masolo, A. Oltramari, and L. Schneider. Sweetening ontologies
with DOLCE. In 13th International Conference on Knowledge Engineering and Knowledge
Management (EKAW02), volume 2473 of Lecture Notes in Computer Science, page 166 ff,
Siguenza, Spain, Oct. 1--4, 2002.
 Pease, A., Niles, I., and Li, J. 2002. The Suggested Upper Merged Ontology: A Large
Ontology for the Semantic Web and its Applications. In Working Notes of the AAAI-2002
Workshop on Ontologies and the Semantic Web, Edmonton, Canada, July 28-August 1, 2002.
 Westenthaler, Arora, Maass, Behrendt (2005) Implementation of Knowledge Content Objects,
Knowledge Content Carrier Architecture and Knowledge Content Transfer Protocol.
METOKIS Deliverable D16, 66 pages.
Mag. Veronika Hornung-Prähauser, MAS
Salzburg Research Forschungsgesellschaft
Jakob Haringerstrasse 5/3
5020 Salzburg AUSTRIA