=Paper=
{{Paper
|id=Vol-273/paper-1
|storemode=property
|title=Ontology Maturing: a Collaborative Web 2.0 Approach to Ontology Engineering
|pdfUrl=https://ceur-ws.org/Vol-273/paper_14.pdf
|volume=Vol-273
}}
==Ontology Maturing: a Collaborative Web 2.0 Approach to Ontology Engineering==
https://ceur-ws.org/Vol-273/paper_14.pdf
Ontology Maturing:
a Collaborative Web 2.0 Approach to Ontology Engineering
Simone Braun Andreas Schmidt Andreas Walter
FZI Research Center for FZI Research Center for FZI Research Center for
Information Technologies Information Technologies Information Technologies
Karlsruhe, Germany Karlsruhe, Germany Karlsruhe, Germany
braun@fzi.de aschmidt@fzi.de awalter@fzi.de
ABSTRACT closer look at the processes of creating and especially main-
Most of the current methodologies for building ontologies taining such domain-specific ontologies. In real-world set-
rely on specialized knowledge engineers. This is in contrast tings these issues are crucial to fulfill the users’ needs and
to real-world settings, where the need for maintenance of do- currently insufficiently dealt with.
main specific ontologies emerges in the daily work of users. It is usually acknowledged that ontologies are shared un-
But in order to allow for participatory ontology engineering, derstandings of a particular domain that have to be con-
we need to have a more realistic conceptual model of how structed within social processes among the stakeholders.
ontologies develop in the real world. We introduce the on- However, current methods and tools do not empower the
tology maturing processes which is based on the insight that users to actually carry out these negotiation processes em-
ontology engineering is a collaborative informal learning pro- bedded in their work.
cess and for which we analyze characteristic evolution steps This is mainly due to lopsided and partially naive perspec-
and triggers that have users engage in ontology engineering tives on ontology engineering. One perspective requires on-
within their everyday work processes. This model integrates tology engineering experts who moderate the ontology cre-
tagging and folksonomies with formal ontologies and shows ation processes and on whom users will depend indefinitely.
maturing pathways between them. As implementations of In the other perspective, it is assumed that users become
this model, we present two case studies and the correspond- ontology modeling experts right away. Presupposing an on-
ing tools. The first is about image-based ontology engineer- tology engineering expert, we have to face two problems:
ing (introducing so-called imagenotions), the second about First, involving ontology engineering specialists is expensive
ontology-enabled social bookmarking (SOBOLEO). Both of and second, ontologies produced by modeling experts in-
them are inspired by lightweight Web 2.0 approaches and stead of domain experts can contain errors caused by the
allow for realtime collaboration. insufficient domain knowledge and experience of knowledge
engineers [2]. Leaving the complete, challenging and com-
plex task of ontology modeling to the domain experts on the
Categories and Subject Descriptors other hand is usually not an option either! Often users are
H.1.m [Miscellaneous]; H.5.3 [Group and Organization not motivated to invest the effort because they are concerned
Interfaces]; I.2.4 [Knowledge Representation Form- with their work processes and regard ontology modeling in
alisms and Methods]: Semantic networks; K.4.3 [Org- its traditional form as an overhead. The main reason there-
anizational Impacts]: Computer-supported collaborative fore is the fact that the time lag between the emergence of
work concepts and their inclusion in ontologies is far too big for
ontologies to be useful [14]: concepts are already becoming
General Terms obsolete by the time they are entering the ontology.
Indeed, users are almost constantly constructing and ne-
Design, Human Factors, Theory gotiating shared meaning in collaboration with others by
augmenting and evolving a community vocabulary. The
Keywords main challenge is then how to leverage this implicit and
Ontology engineering, ontology maturing, Web 2.0, tagging, informal ontology building for the explicit formal models
social software, work integration, collaboration needed for semantic approaches. What we actually need in
order to cope with this challenge are two things: First, on
the conceptual level: a more realistic and work-integrated
1. INTRODUCTION view of how ontologies actually are or can be created and
Within state-of-the-art semantic approaches, ontologies second, on the technical level: tools supporting interweaving
have emerged as the key to enable more advanced technolog- working and ontology engineering activities and the associ-
ical support for end users and their work processes, which ated social negotiation process.
particularly applies to knowledge work. However, current This paper is organized as follows: Section 2 discusses re-
research and development concentrates more on what we lated work. In section 3, we present our notion of ontology
can do as soon as we have ontologies—rather than having a maturing as a conceptual model and analyze the motiva-
Copyright is held by the author/owner(s). tion for users, to participate on such collaborative maturing
WWW2007, May 8–12, 2007, Banff, Canada. processes. We present two case studies in section 4 and 5—
.
they show how this conceptual model in combination with ging systems using the example of Del.icio.us [5] and Flickr
lightweight and collaborative Web 2.0 tools enables and fos- [8]. Golder & Huberman [9] observed that tags used for
ters the maturing process. Section 6 concludes. an individual resource stabilize over time. Marlow et al.
[20] showed that the tag vocabularies of socially connected
users have a bigger overlap than those of randomly selected
2. RELATED WORK users. Both might be traced back to issues of imitation and
There are several proposed methodologies for the process an upcoming shared knowledge. Sen et al. further inves-
of ontology engineering that require ontology engineers as tigated the community impact on personal vocabulary and
moderators. These methodologies define how to support tagging behavior [25]. They explored different forms of se-
the ontology lifecycle from development, via evaluation and lecting tags used in the community. They observed that
maintenance, to further evolution. An overview is given in users apply (“borrow”) an already used tag more often in
[7] and [10]. They do not allow for the work-embedded and case tags of community members compared to when no tags
collaborative engineering of ontologies are displayed. Our model in the next chapter considers these
A more “human-centered approach” is taken by [17] with observations as a change from ideas to consolidation in com-
the Human-Centered Ontology Engineering Methodology munities.
(HCOME). Kotis et al. [17] view ontology development as
a dynamic process and focus particularly on ontology evolu-
tion. They assume a decentralized engineering model where
3. THE ONTOLOGY MATURING PROCESS
everyone first formalizes her own ontology and shares it in In the following, we will present the steps of our model
a further step within the community. There, the individ- and the motivation and triggers for user to participate on
ual ontologies are merged or further developed. However, the ontology maturing progress.
findings in [1] (based on action theory) suggest that collab-
oration plays a more important role before we have formal-
3.1 Observations about ontologies
ized (individual) ontologies. So we think that the HCOME At first, we will collect some important observations about
methodology can benefit from incorporating the notion of ontologies, based on our experiences, that help in developing
different maturity levels. This methodology does not sup- a more appropriate model for building ontologies:
port our goal of embedding ontology engineering of ontolo-
gies in work processes. • Ontology building as a learning process. Key is
Editing tools like Protégé [22] or KAON OIModeler [19] the insight that building ontologies is not just about
are commonly used for ontology building ([21] and [6] dis- eliciting knowledge and formalizing it according to a
cuss them in detail). Most existing ontology editors are particular formalism. Rather this construction process
standalone desktop applications. In addition, they con- itself is a learning process in which the involved indi-
sider ontology construction as an isolated and detached task viduals deepen their understanding of the real world
where ontology engineering experts explicitly sit together and of an (appropriate) vocabulary to describe it. This
with knowledge workers to model the ontologies. They do is especially true if we consider emerging ideas: it is
not provide a collaborative environment or support collabo- simply not possible to integrate these into an ontology
ration only in a restricted way (except for KAON). Like the straight away.
above mentioned methodologies, these tools are not geared • Formality and complexity of use as barrier. Re-
towards knowledge workers and their work processes. lated to that learning process, we should also consider
Wiki systems consider the aspects of collaboration and the lessons learnt from the success of Web 2.0 ap-
can support the early phases of ontology construction. Se- proaches. Web 2.0 approaches empower the individual
mantic wiki systems1 try to extend the traditional wikis with to take part in community activities by lowering the
semantic web technologies (e.g. Semantic MediaWiki [28], barriers: informal, lightweight, easy-to-use, and easy-
OntoWiki [15] or IkeWiki [23]). These systems help users to-understand. Tagging as an organizing paradigm has
in creating definitions, e.g. beginning with informal texts. replaced folder hierarchies. As a conclusion, we should
Because of discussion pages and versioning for each article acknowledge that the degree of formality poses a bar-
they are suitable for complex coordination and consolida- rier. And that the compelling simplicity of Web 2.0
tion processes, but they are usually more time-consuming applications is part of their success model.
and lack work process integration.
Collaborative tagging systems form a third application • Continuous evolution in work processes. Ontol-
group that is getting more and more popular on the web. ogy building is usually not supposed to be a one-time
Their lightweight approach allows users to easily assign key- activity of an expert committee, but rather a sustain-
words to various contents. The social bookmarking systems able process of continuous evolution. But if we can-
BibSonomy ([16]) and Del.icio.us [5] further allow grouping not assume that everything is on the same formality
of tags underneath a super-tag. However, they do not define level, we need to acknowledge that in a living ontol-
a clear concept structure and this functionality is very re- ogy there are most likely concepts not clearly defined
stricted. For instance these applications do not display the yet. They underlie a process of continuous evolution
tree structure formed by these relations and the structure where ideas and understanding emerge implicitly in
on top of tags is not shared. daily work and mature only gradually through the in-
There were first attempts to analyze occurrence patterns teraction with others to explicit formal and shared con-
and in particular the usage of tags within collaborative tag- ceptualizations.
1
See website of the SemanticWiki Interest Group for up-
dated list of existing semantic wiki systems [12]
new concept ideas common formal
tags terminology lightweight
ontology
heavy-weight
ontology
Emergence Consolidation Axiomat-
Formalization
of ideas in Communities ization
Figure 1: Ontology Maturing Process
3.2 The Model of Ontology Maturing 3. Formalization. Within the third phase, concepts are
These observations are similar to those made in [24] about organized into relations - both taxonomical (hierarchi-
how new ideas develop in the context of knowledge man- cal) and ad hoc relations. For instance, we want to
agement and e-learning to become reusable training ma- receive recommendations from the system and we be-
terial. This development process was described with the come aware of different abstraction levels (e.g. broader
metaphor of maturing and structured into five phases as and narrower) that we need to have in order to find
the so-called knowledge maturing process. This process is e.g. spaghetti and bigoli (thicker spaghetti) as noodles
viewed as a macro model for interconnected individual learn- and vice versa. The results are lightweight ontologies
ing processes. Based on this process model, we have come that rely primarily on inferencing based on subconcept
up with a similar phase model that identifies characteristical relations.
maturing transitions in collaboratively developing a shared
ontology (see Figure 1): 4. Axiomatization. The last phase of our model cap-
tures more domain semantics by adding background
1. Emergence of ideas. In this initial transition, new knowledge for improving inferencing processes, e.g. for
concept ideas are introduced which are rather ad hoc query answering. This step requires a high level of
and not well-defined. They are personal “utterances” competence in logical formalism so that this can usu-
which are informally communicated and technically ally only be carried out by domain experts.
typically represented by tags. For instance, while an-
notating or seeking for resources, we recognize that 3.3 Motivation and Triggers for Maturing Ac-
the tag we want to use does not exist or is misspelled. tivities
Accordingly, we introduce a new tag or correct the ex- This model as such only describes characteristic degrees
isting one without further reflecting. of formality of ontology elements from a user-driven point of
view. But it does not explain yet, how this integrates with
2. Consolidation in Communities. Through reuse actual work processes.
and adaption of concept symbols (tags) of others, a When and how do users engage in maturing activities?
shared vocabulary emerges within a community. When What triggers the maturing of concepts and other ontology
comparing currently envisioned tags with previously elements? And why do we believe that users are motivated
used ones or with tags from other people assigned to to participate in this collaborative maturing effort?
the same resource, we discover similarities and differ- Starting point for our analysis is the observation that on-
ences that allow for creating concepts from tags. For tologies are mainly used and deliver their main benefit for
instance, we realize that we can improve our search by information seeking or distribution activities. This includes
using a synonymous term. We establish a link between users wanting to:
these terms in our understanding and thus can merge
synonyms into concepts. Moreover, preferred labels • retrieve appropriate content
also often develop in the same way [9]. It crystallizes
a common terminology still without formal semantics. • promote their content so that it is found by others
• enhance their work performance (getting answers in- • Missing guidance to related information. Es-
stead of searching) pecially the orientation phase of information seeking
processes (like [18] described) requires guidance to find
• be empowered by being able to state their opinions related fields. As soon as implicit links are discovered,
they can be added to the ontology. An increasing num-
This coincides with motivations for tagging resources in ber of links (at the latest) leads to defining new prop-
[20]: Future retrieval, Contribution and Sharing, Attract At- erties.
tention, Play and Competition, Self Presentation, and Opin-
ion Expression. 3.4 Tool support
The embedding of ontology maturing into users work proc-
In the following two sections, we present two approaches
esses also requires to have a look at the micro level of con-
that support the collaborative maturing process of ontolo-
crete triggers initiating these activities. These triggers are
gies based on our introduced model and the relevant triggers.
important to consider because they are the technical inte-
Both provide tool support in work integrated scenarios.
gration points where we need to make the transition between
The first case study is rooted in the domain of semantic
working and ontology maturing possible.
image retrieval. Here, the focus lies on the consolidation
If we consider the retrieval of appropriate content, we
of image descriptions in communities with the help of vi-
can identify following typical triggers which are closely con-
sualized concepts and relations. The case stuy introduces a
nected to semantic shortcomings of tagging systems (cf. [9],
new method called imagenotions. This method allows the
[13]):
collaborative maturing of unstructured tags to commonly
accepted concepts.
• (Mis-)Spelling. The most obvious problem is that The second approach also provides consolidation capa-
tags were simply misspelled or that there are typo- bilities, but in particular concentrates on how to support
graphical differences, for instance because of occurring the formalization based on consolidated concepts. The key
plurals, abbreviations or compound words. Even inex- idea here is to couple social bookmarking approaches with
perienced users tend to correct such issues if possible. a lightweight ontology editor. Its application domain is in-
formal learning and knowledge management support in in-
• Synonymy. Documents are not found with the first
terdisciplinary application-oriented research.
keyword, but later under a synonym. Similar is the
issue for the author of a resource trying to promote
its usage: she has to use many synonyms in order to 4. IMAGINATION
ensure that other users will find it later on. This typi- The IMAGINATION EU project2 provides image-based
cally leads users to abstract from keywords to concepts navigation for digital cultural and scientific resources. Users
with several synonyms (possibly also in multiple lan- can click on parts of an image to find other interesting im-
guages). ages to a given context. A click on an image part auto-
matically generates a semantic query. This feature requires
• Multilingualism. Tags are always in one language. semantic metadata. In IMAGINATION, semantic metadata
This requires that content owners, especially in Eu- are automatically generated by combining text-mining, ob-
rope with many different languages, have to describe ject recognition and object identification algorithms exploit-
resources with many tags in different languages. ing domain ontologies. The metadata can be validated, cor-
rected and extended manually if needed.
• Homonymy. A tag can have different meanings, so
called homonyms. This leads to search results with low 4.1 Requirements
precision, as all resources that are relevant to these Based on our experience in implementing commercial im-
different meanings are annotated with the same tag. age search systems3 , current metadata about the content
Fixing this issue is more demanding as it requires to of images is largely based on the unstructured and non-
differentiate existing tags into different concepts. Its semantic tagging paradigm (even before the popularity of
effects can only be realized if resources are reindexed, Web 2.0)—often called also keyword or label and embed-
which often is not manageable. ded into IPTC attributes [4]. This tag-based annotation of
images has many of the shortcomings mentioned as triggers
• Missing concepts. During annotation activities, users for entering ontology engineering activities in the previous
discover that a topic they want to describe is not yet section. Previous attempts to solve the problem with the
covered. This leads to additions to the ontology. help of thesauri have largely failed because thesauri were
not accepted by the users and content providers due to their
• Mismatch of abstraction level. A typical search incompleteness and/or complexity.
problem in complex domains is that search terms are These experiences can be directly fed into requirements
specified either too broad or too narrow. This prob- for the use of ontologies in the area of image search and
lem, also known as the “basic level phenomenon” [27], navigation: (1) the ontology (and therefore semantic anno-
can be traced back to different intentions and exper- tations using it) must be comprehensible for the users and
tise levels of the users. For instance, one user tags a (2) an ontology must cover the image repository completely.
resource on the basic level with “spaghetti”, another The challenge is that image repositories can change rapidly
with “noodles” and a third differentiates “spaghetti” and consequently the ontology must be adapted regularly.
from “bigoli” (thicker spaghetti) and “vermicelli” (thin-
2
ner spaghetti). Thus, later success in finding leads to http://www.imagination-project.org
3
the adding of hierarchical relationships. See e.g. www.fotomarktplatz.de
4.2 Ontology Development Methodology Based 3. Formalization. In this step, taxonomical (“is-a”) and
on Imagenotions ad-hoc relations are specified among imagenotions.
Based on the ontology maturing process, we can develop
a solution for these problems. In terms of the previously in- After step 2, the quality of image search already increases
troduced ontology maturing process we concentrate on the significantly, as the problems with synonyms and homonyms
first three steps from the emergence of ideas up to formaliza- do not appear anymore. Moreover, it is easy to see that all
tion. What distinguishes our methodology from the usual annotated images automatically benefit from the maturing
ontology development methodologies is the strong emphasis imagenotions. E.g. adding a new tag to an imagenotions
on collaborative ontology development in the consolidation- automatically allows users to find all images annotated with
in-communities phase. This is motivated by the success of that imagenotion using the new tag. In addition, the out-
collaborative tagging in Web 2.0 projects. come of step 3 also allows requests for related images based
The basis of our ontology formalism is a concept we call on the current context.
imagenotions. An imagenotion graphically represents a se- Imagenotions are useful to collaboratively build an on-
mantic notion through an image, or a set of images. In tology supporting manual annotation and semantic search.
addition, similarly to many existing ontology formalism, it However, to fulfill the requirement of IMAGINATION for
is also possible to associate tags with an imagenotion in automatic annotation, a classical formal ontology is needed
different languages (such as English or German). For each that can be exploited by text-mining and object identifica-
language, one of these tags is selected as the main label of tion algorithms. This last axiomatization step is not yet di-
the imagenotion. The other tags are termed synonyms. rectly supported by our methodology, it is subject of future
Instead of tags, images are annotated with imagenotions. research. Nevertheless, it is easy to see that a conversion of
It is easy to see the advantage: all the shortcomings of tag- an imagenotion ontology to a standard ontology formalism
ging approaches are solved using imagenotions—it is easy (e.g. OWL) is possible. The only missing information is
to find images using search terms in any language. In other whether an imagenotion should be modelled as a concept,
words, we provide semantic search instead of full-text search. instance or a relation in the target ontology formalism.
In the terminology of classical ontologies, imagenotions
are usually instances, but they may also correspond to con-
4.3 Tool Support
cepts or relations. There are two major advantages of using Currently we implement a web-based tool that allows the
imagenotions over the classical ontology constructs: creation of new imagenotions and the editing of available im-
agenotions. This tool supports all three steps of our method-
1. The distinctions between concepts, instances and re- ology. It can be easily invoked during semantic search or
lations are hard to understand for most users. In our when uploading new images into an image repository: e.g.
mind, notions play the role of an instance, a concept or it is fully integrated into the user’s workflow.
a relation, depending on the actual context. This fact We now demonstrate some functionality of the tool in
is acknowledged by many ontology formalisms that al- terms of the steps of our development methodology.
low metamodeling. Using imagenotions, users do not
need to understand this somewhat artificial separation 4.3.1 Step 1: Emergence of Ideas
of notions. Figure 2 shows an example for the emergence of ideas.
Let us assume that a content owner has new images about
2. Because imagenotions are associated with images, they
elephants. The imagenotion “elephant” was so far not avail-
are meaningful internationally as an image has the
able. Therefore, she creates a new imagenotion, adds an
same meaning in different languages.
image or part of an image that shows elephants and starts
The goal of our methodology is to guide the process of describing the new imagenotion with more details. She uses
creating an ontology of imagenotions. The main steps of English as spoken language. As synonyms, she enters “ele-
this methodology is based on the ontology maturing process phantidae” and “tusker”. Instead of tagging the new images
model: that show elephants with these words, she can use the new
imagenotion—she just pulls this imagenotion over the new
1. Emergence of Ideas. In this step, new imagenotions are images via drag and drop.
created. Already this step can become collaborative,
as users can jointly collect the tags describing imageno-
tions, and select the most representative images for an
imagenotion. Collaborative editing is especially use-
ful in a multi-lingual environment where it cannot be
expected that any individual user speaks all required
languages.
2. Consolidation in Communities. Because it is so easy to
create new imagenotions, it cannot be avoided that for
the same semantic notion initially many imagenotions
are created (synonyms, also in different languages) or
that an imagenotion represents more than one seman-
tic notion (homonyms). In this step, these problems
should be solved by merging synonymous imageno-
tions, and by splitting imagenotions representing more Figure 2: Editing an imagenotion with the No-
than one notion. tionEditor tool
Figure 3: Spliting homononyms in different im- Figure 5: Adding relations to imagenotions
agenotions
Figure 6: Visual search refinement
Figure 4: Creating hierarchies in imagenotions
imagenotion “mammal” and adds it by drag and drop to the
The front end of the image search system also allows users imagenotion of elephants to state that “every elephant is a
to add missing synonyms to already defined imagenotions in mammal”. It is also possible to define other kinds of rela-
order to achieve better search results. The label and syn- tions. For instance, in Figure 5 somone defined the relations
onyms of imagenotions are editable in different languages. “has baby” and used this relation between the imagenotions
All images described with imagenotions automatically ben- “Angelina Jolie” and her baby “Shiloh” as well as her son
efit from all added information. In other words, the com- “Maddox Chivan”.
munity improves the quality of the image descriptions by
improving the quality of the corresponding imagenotions. 4.4 Application for Semantic Search
In addition to improving search results, imagenotions also
4.3.2 Step 2: Consolidation in Communities allow a new principle of visual search refinement based on
Consolidation solves the problem of homonyms and differ- formal knowledge. Users start by typing some search terms.
ent languages. For instance, the word “tusker” is a homonym Our system proposes imagenotions containing these search
because (among others) it is also the brand of a Kenyan beer terms as label or synonym. Furthermore, our system also
brewery (see Fig. 3). The first user noticing this homonym proposes images and search refinements based on the current
can solve the problem: she splits the imagenotion tusker in context. These proposals are visualized with imagenotions
two new imagenotions—one with an image of an elephant and determined using the relations from the available im-
for the meaning “elephant”, and another with the label of agenotion ontology. Refining of a search request is possible
the Kenyan brewery for the meaning of the “Kenyan beer by just clicking on the desired imagenotion.
brand”. Figure 6 shows such an example. The current search con-
In the case when someone detects two imagenotions that text is “images that contain the actors Angelina Jolie and
have the same meaning (possibly in different languages), she Brad Pitt” and the user already selected a matching image.
can merge them together. The system proposes two related groups of images: one for
the film “Mr. and Mrs. Smith” and one for their common
4.3.3 Step 3: Formalization baby. These proposals are visualized using the correspond-
Figure 4 shows an example for formalization—the creation ing image notions.
of hierarchies. In this example two new imagenotions for two
sorts of elephants are created: the “African elephant” and
the “Asian elephant” imagenotions. The trigger for creating
5. IM WISSENSNETZ – IN THE KNOWL-
hierarchies may be the description of images in more details. EDGE WEB
As elephants are mammals, one could also add broader Our second use case is taken from the German research
imagenotions (the “is-a” relation): the user searches for the project “Im Wissensnetz – Vernetzte Informationsprozesse
in Forschungsverbünden”4 (“In the Knowledge Web – Net-
worked Information Processes in Research Associations”),
which aims to support efficient interdisciplinary knowledge-
added processes within e-Science. Research is likely to be
the most knowledge-intensive environment. An empirical
analysis of existing (cooperation-)processes, information and
knowledge exchanges, and instruments for the preservation
of knowledge accomplished in the application domain “rapid
prototyping” revealed that scientific work is characterized by
high variability, dynamics and unpredictability as well as by
high significance of social interactions and communication.
5.1 Requirements
Linking people with individual expert knowledge and con-
tents from various disciplines like plastics, ceramics, and me-
chanical engineering is one of the most important challenges.
For instance, one major problem is searching and retrieving
adequate contemporary resources. This process is very te-
dious. The users have to access many various data sources
with different interfaces, but also the internet with common
search engines like Google. Figure 7: Tagging a web resource with concepts.
In the area of plastics and their market these high dynam-
ics are particularly obvious. New materials or new forms
of existing ones frequently enter the market; brand names rial, they can pick it up into the shared bookmark collection
and manufacturers are permanently changing and hardly and annotate it with concepts from the vocabulary, e.g. the
trackable—attributes of a chemical substance retrievable us- specific plastic. If a needed concept does not exist in the
ing its brand name today, are very hard to find once it’s sold shared vocabulary (e.g. in the case of a new material) or
under a different label. There is also no general up to date is not suitable (e.g. when the brand name changed), the
database which list manufacturer and brand names of cur- users can modify an existing concept or add arbitrary tags.
rently available forms of plastics. Thus, the users rely on These new tags are automatically collected as “prototypical
search engines like Google in order to find the wanted mate- concepts” within the vocabulary. The users can consolidate
rial of which only the old brand or chemical name is known. them later by abstracting into concepts and placing them
However, common search engines provide many irrelevant within the ontology. In this way, we allow that new con-
results because of their missing focus on the domain. For cept ideas and tags are gathered seamlessly when they are
instance, when looking for “nylon” you receive lots of results occurring and that users can define concepts in a free and
for stockings. informal manner without the usual modeling overhead.
At this point, using annotation and retrieval tools can For structuring the concepts we concentrate on taxonomic
help; e.g. when a colleague already found the new brand relations because they are easy understandable for non-mo-
name of a product and tagged it with the old one you are delling experts. Going beyond common tagging systems,
looking for. With such tools being further semantically en- where the users can only bundle concepts underneath an-
riched with background knowledge and domain ontologies, it other, the users can organize their concepts within our ap-
is possible to find out the search context and thus to extend proach in a shared structure according to the SKOS Core Vo-
or refine the search in order to reduce irrelevant results and cabulary [3] in which the concepts are connected by “broader”,
to guide the user. However, this also requires that (1) the “narrower” and “related” links. The users can further spec-
users collaboratively build up and maintain a shared under- ify one preferred label and a number of alternative labels or
standing and terminology, (2) these activities are embedded synonyms (e.g. former brand names) as well as a textual
into their information seeking activities and (3) this under- description for each concept.
standing is expressed formally enough to enable ontology- The users within the community share and maintain one
based query refinement. taxonomy and one collection of bookmarks collaboratively.
Everyone has the right for editing and modifications follow-
5.2 Ontology Development Methodology Based ing the wiki paradigm of self-regulation.
on Semantic Social Bookmarking In this way, we can support in particular the second and
third transition phase of the ontology maturing process.
In order to overcome these problems, we provide a solu-
tion that combines collaborative collecting and sharing of 5.3 Tool support with SOBOLEO
web resources (bookmarks) with collaborative development
We are developing SOBOLEO (Social Bookmarking and
of a shared vocabulary. This vocabulary is used to organize
Lightweight Engineering of Ontologies) in order to satisfy
the bookmarks. That means, collected bookmarks can be
these requirements. SOBOLEO’s goal is to support knowl-
annotated with concepts from the vocabulary (see Figure
edge workers working together in one domain in developing a
7).
shared vocabulary and a shared collection of relevant web re-
If users find a web resource, e.g. the manufacturer’s web-
sources with a lightweight ontology editor and an ontology-
site for a specific plastic or an article about a new mate-
enabled social bookmarking system.
4
http://www.im-wissensnetz.de/
Figure 8: Ontology editor user interface. Figure 9: Searching the bookmark collection.
SOBOLEO is based on AJAX technology using the Google • Annotation. The annotation interface opens a pop
Web Toolkit [11] and works in most current browsers—thus up (see Figure 7) in which users can enter title and url
does not require any local installation. It consists of four ap- of the web resource they want to add to the collection.
plication parts: an editor for the modification of the shared Users can further specify concepts from the taxonomy,
taxonomy, a tool for the annotation of web resources, a se- which is also supported by auto completion, or they
mantic search engine for the annotated web resources and can add new tags to annotate the resource. If the
a taxonomy browser for navigating the taxonomy and the collection already contains the resource, the existing
bookmark collection. annotations are displayed and can also be edited.
• Search. The semantic based search engine allows for
• Editing. The editor interface (see Figure 8), built searching and retrieving resources within the shared
up tripartite, displays a tree view of the taxonomy on bookmark collection (see Figure 9). Users can enter
the left hand side of the screen. It shows the con- concept labels from the taxonomy or arbitrary search
cepts with their preferred labels and their narrower terms. The engine presents all resources either anno-
and broader relations. Informal, not yet consolidated tated with the identified concepts and their narrower
tags are collected in the special branch of “prototypi- ones or containing the input terms within the page
cal concepts”. When a concept is selected in the tree content. The interface lists the resources with their
view, its details are displayed in the center part of title linking to the original page, with annotated con-
the screen. Here, the users can edit the preferred and cepts, a short excerpt of the page content highlighting
alternative labels, the description, and the narrower, the search terms, and the exact url. It further pro-
broader and related relations between concepts, which vides query refinement and relaxation proposals. Via
is further supported through auto completion of enti- each result’s edit link, users can modify or remove the
ties in the taxonomy. The right hand side of the screen annotations of a web resource.
provides a chat panel and allows having a conversation
with other people editing the same ontology at the • Browsing. With the browsing interface users can
same time. The chat panel is also used for displaying navigate through the taxonomy and associated book-
changes made to the taxonomy. The system automat- mark documents (see Figure 10). Starting from the
ically generates a chat message that details who did root concepts, the users can click through the taxon-
which change. Changes done to the ontology by one omy concepts. On top, the users see the currently
user are visible almost instantaneously on all machines selected concept with its preferred and alternative la-
without requiring any intervention by the users. Up- bels and its description. Additionally, all its broader,
dates of other users are also immediately reflected in narrower and related concepts are displayed as links
the center screen part showing the details of the cur- for further navigation. Underneath the concept de-
rently selected resource—particularly important if two tails there is a list of all resources which are annotated
users are editing one concept at the same time. with the currently selected concept or with one of its
for the collaborative creation of required domain ontologies.
SOBOLEO allows social bookmarking with ontologies in the
project Im Wissensnetz.
Our next work steps are evaluations and refinements of
our model for the ontology maturing processes in these pro-
jects. We are mainly interested in evaluating which fur-
ther functionalities are needed and wanted by the users, e.g.
whether and when it is necessary to introduce big numbers
of relations between the available concepts. Likewise, for
the sake of simplicity, we are not considering aspects of on-
tology versioning at the moment (here, we want to refer to
the works of [26]). Our main evaluation target is to analyze
if the shared ontologies, developed based on our tools and
models, converge to a stable and common accepted under-
standing for the available concepts over time. If not, our
models require further support for agreement finding (e.g.
like wiki) or personal ontologies.
7. ACKNOWLEDGMENTS
This work was co-funded by the European Commission
within the project IMAGINATION, by the German Federal
Ministry for Education and Research within the project Im
Wissensnetz and by German Federal Ministry of Economic
and Technology within the project KSIunderground.
8. ADDITIONAL AUTHORS
Additional authors: Gabor Nagypal (disy Informations-
Systeme, Karlsruhe, Germany, email: nagypal@disy.net)
Figure 10: Browsing the taxonomy and associated and Valentin Zacharias (FZI Research Center for Informa-
bookmarks. tion Technologies, email: zach@fzi.de).
9. REFERENCES
narrower concepts. These resources are further ranked [1] H. Allert, H. Markannen, and C. Richter. Rethinking
by their date they were collected, thus the newest re- the Use of Ontologies in Learning. In M. Memmel and
sources appear upmost. D. Burgos, editors, Proceedings of the 2nd
International Workshop on Learner-Oriented
6. CONCLUSIONS Knowledge Management and KM-Oriented Learning
(LOKMOL 06), in conjunction with the First
We will only ever achieve sustainable ontology-based sys- European Conference on Technology-Enhanced
tems by embedding the task of building and maintaining Learning (ECTEL 06), pages 115–125, October 2006.
ontologies into everyday work processes, enabling domain
[2] K. Barker, V. K. Chaudhri, S. Y. Chaw, P. Clark,
experts to do it without the help of knowledge engineers and
J. Fan, D. Israel, S. Mishra, B. W. Porter, P. Romero,
by making it truly collaborative. We also have to acknowl-
D. Tecuci, and P. Z. Yeh. A question-answering
edge that ontologies cannot be formalized from scratch, but
system for AP Chemistry: Assessing KR&R
rather continuously evolve in a maturing process from in-
technologies. In Proceedings of the Ninth International
formal tags to formal taxonomy hierarchies for which the
Conference on Principles of Knowledge Representation
ontology maturing process was presented.
and Reasoning, pages 488–497, 2004.
Therefore, our model for ontology maturing offers four
[3] D. Brickley and A. Miles. SKOS Core Vocabulary
different steps. It allows for the emergence of ideas from
Specification. W3C working draft, W3C, November
each individual and the consolidation in communities for a
2005.
common terminology. Then, in the third step of our model,
relations help in creating formal lightweight ontologies. Fi- [4] I. P. T. Council. The IPTC Standard, 2007.
nally, the fourth step could allow axiomatization. Such a [5] Del.icio.us. Del.icio.us. http://del.icio.us/, 2007.
maturing view on ontology engineering can overcome the (accessed 2007-01-25).
problem of conceptual dynamics (e.g. the problem of the [6] M. Denny. Ontology Tools Survey, Revisited.
time lag between emergence of topics and their inclusion in Technical report, XML.com, July 2004.
an ontology). [7] M. Fernández-López and A. Gómez-Pérez. A survey
In order to support such a maturing process, we presented on methodologies for developing, mainaining,
two lightweight, easy-to-use and work embedded tools that integrating, evaluating and reengineering ontologies.
allow the collaborative maturing of ontologies. Both of them Deliverable 1.4, EU IST Project IST-2000-29243
lower the barriers to ontology editing for non-knowledge for- OntoWeb, 2002.
mulation experts. The project IMAGINATION will use the [8] Flickr. Welcome to Flickr - Photo Sharing.
idea of image based ontolgy maturing with imagenotions http://www.flickr.com/, 2007. (accessed 2007-01-25).
[9] S. Golder and B. A. Huberman. The structure of [19] A. Maedche, B. Motik, and L. Stojanovic. Managing
collaborative tagging systems. Journal of Information multiple and distributed ontologies on the Semantic
Sciences, 32(2):198–208, 2006. Web. The VLDB Journal, 12(4):286–302, November
[10] A. Gómez-Pérez, M. Fernández-López, and O. Corcho. 2003.
Ontological Engineering with examples from the areas [20] C. Marlow, M. Naaman, D. Boyd, and M. Davis.
of Knowledge Management, e-Commerce and the Position Paper, Tagging, Taxonomy, Flickr, Article,
Semantic Web. Advanced Information and Knowledge ToRead. In Collaborative Web Tagging Workshop at
Processing. Springer, 1st edition, 2004. WWW2006, May 2006.
[11] Google Web Toolkit. Google Web Toolkit - Build [21] A. G. Perez, J. Angele, M. F. Lopez,
AJAX apps in the Java language. V. Christophides, A. Stutt, and Y. Sure. A survey on
http://code.google.com/webtoolkit/, 2007. (accessed ontology tools. Deliverable 1.3, EU IST Project
2007-01-25). IST-2000-29243 OntoWeb, 2002.
[12] S. W. I. Group. Semantic Wiki State Of The Art, [22] Protégé. The Protégé Ontology Editor and Knowledge
2007. Acquisition System. http://protege.stanford.edu/,
[13] M. Guy and E. Tonkin. Folksonomies: Tidying up 2007. (accessed 2007-01-25).
tags? D-Lib Magazine, 12(1), January 2006. [23] S. Schaffert. IkeWiki: A Semantic Wiki for
[14] M. Hepp. Possible ontologies: How reality constraints Collaborative Knowledge Management. In 1st
building relevant ontologies. IEEE Internet International Workshop on Semantic Technologies in
Computing, 11(1):90–96, January/February 2007. Collaborative Applications (STICA’06), Manchester,
[15] M. Hepp, D. Bachlechner, and K. Siorpaes. OntoWiki: UK, June 2006.
community-driven ontology engineering and ontology [24] A. Schmidt. Knowledge Maturing and the Continuity
usage based on Wikis. In WikiSym ’06: Proceedings of of Context as a Unifying Concept for Knowledge
the 2006 international symposium on Wikis, pages Management and E-Learning. In Proceedings of
143–144, New York, NY, USA, 2006. ACM Press. I-KNOW ’05, Special Track on Integrating Working
[16] A. Hotho, R. Jschke, C. Schmitz, and G. Stumme. and Learning, July/August 2005.
BibSonomy: A Social Bookmark and Publication [25] S. Sen, S. K. Lam, A. M. Rashid, D. Cosley,
Sharing System. In A. de Moor, S. Polovina, and D. Frankowski, J. Osterhouse, F. M. Harper, and
H. Delugach, editors, Proceedings of the Conceptual J. Riedl. tagging, communities, vocabulary, evolution.
Structures Tool Interoperability Workshop at the 14th In CSCW ’06: Proceedings of the 2006 20th
International Conference on Conceptual Structures, anniversary conference on Computer supported
Aalborg, Denmark, July 2006. Aalborg University cooperative work, pages 181–190, New York, NY, USA,
Press. 2006. ACM Press.
[17] K. Kotis, G. A. Vouros, and J. P. Alonso. HCOME: A [26] L. Stojanovic. Methods and Tools for Ontology
Tool-Supported Methodology for Engineering Living Evolution. PhD thesis, University of Karlsruhe (TH),
Ontologies. In C. Bussler, V. Tannen, and Germany, 2004.
I. Fundulaki, editors, Semantic Web and Databases. [27] J. W. Tanaka and M. Taylor. Object categories and
Second International Workshop - SWDB 2004, volume expertise: Is the basic level in the eye of the beholder?
3372 of LNCS, pages 155–166, Berlin Heidelberg, Cognitve Psychology, 23(3):457–482, July 1991.
Germany, August 2004. Springer-Verlag. [28] M. Völkel, M. Krötzsch, D. Vrandecic, H. Haller, and
[18] C. C. Kuhlthau. Seeking Meaning: A Process R. Studer. Semantic Wikipedia. In Proceedings of the
Approach to Library and Information Services. 15th international conference on World Wide Web
Libraries Unlimited, Westport, CT, 2nd edition (WWW’06), pages 585–594. ACM Press, 2006.
edition, 2004.