We present SOBOLEO, a system for the webbased collaborative engineering of SKOS ontologies and annotation of web resources. SOBOLEO enables the simple creation, extension and maintenance of taxonomies. At the same time, it supports the annotation of web resources with concepts from this taxonomy.
So far, there is no integrated tool, which supports the
collaborative creation of an index of relevant web resources
together with the shared vocabulary used to organize it.
Today’s tools either allow the creation of a shared vocabulary
or to structure information resources according to some fixed
vocabulary. In addition, most tools for editing a shared
vocabulary are tailored to people with training in knowledge
engineering and not for domain experts. That results in a
number of problems for using such tools:
1. Involving engineering experts for knowledge
formulation and modelling causes the creation of a shared
vocabulary to be more expensive [
In order to satisfy these requirement we developed
SOBOLEO. SOBOLEO is the acronym for Social Bookmarking
and Lightweight Engineering of Ontologies. The system’s
goal is to support people working in one domain in the
collaborative development of a shared vocabulary and of a
shared index of relevant web resources. With SOBOLEO it
is possible to create, extend and maintain taxonomies
according to the SKOS Core Vocabulary [
SOBOLEO is based on AJAX technology and works in most current browsers (we tested Internet Explorer 6 and 7, Opera 9, Firefox 1.5 and 2.0). It consists of four major parts: (1) a collaborative real time editor for changing the taxonomy, (2) a tool for the annotation of web resources, (3) a semantic search engine for the annotated web resources, and (4) a taxonomy browser for navigating the taxonomy and the content of the web resources index. Editing the taxonomy or the index of web resources requires authentification. The other areas are open accessible. In the following paragraphs we will describe each of these parts. 3.1
SOBOLEO uses a subset of SKOS [
Entering the editor after succesfull authentification, a user sees the current state of the taxonomy and the changes of other users in realtime.
The interface of the editor (see Figure 1) consists of three main parts. To the left of the screen is a tree display of the taxonomy. It shows the preferred labels of the concepts and the tree formed by the narrower-broader-relations between them. Users can drag’n’drop concepts on other concepts to create new broader relations or drag’n’drop while pressing shift to remove them1. The operations cut, copy and paste (and their shortcuts) are also supported on the concepts in the taxonomy. Buttons below the taxonomy tree allow for the creation and deletion of concepts.
Clicking on a concept displays its details in the center part of the screen. The input fields at the center of the interface can be used to edit the relations between concepts, the labels and the description of the concept. The editing of relations between concepts is supported by auto completion of entities in the taxonomy.
The right part of the screen is reserved for message. The message pane displays automatically generated messages for all changes made to the taxonomy by any user. At the same 1Removing broader relations in this way is currently not working in Firefox time, the message pane acts as chat window and shows all exchanged chat messages. An input field below allows the user to send chat messages to other users currently editing the taxonomy. 3.2
Clicking on the annotate button anywhere in SOBOLEO opens the annotation interface (see Figure 2) as a pop up. In this interface, users can enter the url and the title of the document they want to add to the index. In another input field they can add a number of concepts from the taxomomy or arbitary tags to annotate this new resource. This topic field offers auto completion of entities in the taxonomy2 to encourage users to apply these concepts. Tags not previously known to the system can be used as well and will appear in the taxonomy editor as “prototypical concepts”; users can consolidate and move them within the taxonomy later. In this way we allow for the seamless gathering of new concept ideas when they are occuring.
A better way to use the annotation tool is, however, to drag the annotate button on to the browser bookmarks, where it will work as a bookmarklet. Whenever a user visits a site she deems interesting for the index, she can click on the bookmarklet and it will open the annotation interface with url and title already filled in based on the page that is open in the browser.
The SOBOLEO system automatically sends out a crawler to any newly added url to fetch and index the contents of the page3. The contents of the page together with any concepts used to annotate it gets added to the index.
Whenever the annotate interface is opened for a particular url, the system checks whether there is already an annotation. If so, the annotation is displayed and can also be 2Based on the preferred label of the concepts 3Please note that SOBOLEO cannot add the content of sites that use frames or require authentification—in these cases only the title, the url and the annotated concepts will be saved edited. In the current system one instance of SOBOLEO has one annotation per url. This annotation consists of a title and an arbitrary numbers of concepts. Annotations are maintained collaboratively. 3.3
SOBOLEO includes a semantic search engine supporting search and retrieval of the indexed web resources. The SOBOLEO search engine is used by typing search terms into a text field—similar to common internet search engines. The entered search string, however, is analyzed for occurrances of concept labels from the taxonomy. An occurrance of a label of a concept in a search string is understood as a reference to a concept. If references to concepts can be found, the SOBOLEO search engine searches for web resources annotated with these concepts or narrower ones (related concepts are currently not regarded). The results from the semantic search are combined with the result from a full text search in the contents of all annotated pages.
On the result page (see Figure 3) the SOBOLEO search engine gives feedback on which concepts it understood the query to be referencing. Depending on the search string, the results and the taxonomy, the system may also propose a number of query refinements or relaxations. The interface lists all found resources with their title linking to the original page, with annotated concepts, a short excerpt of the page content highlighting the search terms and the exact url. The edit links for each result allow to quickly modify or remove the annotation of a web resource. The annotation pop up interface is used (see Figure 2). 3.4
With the browsing interface users can navigate through the taxonomy and the content of the web resources index (see Figure 4). Starting from the root concepts, the users can click through the taxonomy concepts. On top, the users see the currently selected concept with its preferred and alternative labels and its description. Additionally, all its broader, narrower and related concepts are displayed as links for further navigation. Underneath the concept details there is a list of all resources which are annotated with the currently selected concept or with one of its narrower concepts. These resources are further ranked by their date they were collected, thus the newest resources appear upmost.
The server side of SOBOLEO is implemted in Java on
top of the Apache Tomcat [
A large number of other approaches for the creation of
structured vocabulary and the building of a shared web
index exist. We consider the semantic wikis (i.e. [
Compared to Semantic Wikis it is immediately obvious that SOBOLEO is much more spezialized. Semantic Wikis are general purpose knowledge formulation tools—SOBOLEO is deliberately restricted to a simple subset of SKOS. Because of this restriction SOBOLEO can offer more specialized user interfaces and search facilities making it faster, simpler and more effective for the creation of a shared web index together with the taxonomy used to organize it. SOBOLEO’s editor also allows for a closer collaboration among the editors and virtually eliminates the chance for edit collisions. On the other hand, however, it is unclear whether SOBOLEO’s interface could scale to knowledge models of a size that rivals Wikipedia.
Compared to state-of-the-art social web annotation tools
like del.icio.us, SOBOLEO offers more ways to structure
the shared vocabulary that is used for the annotation. The
state-of-the-art structure in such tools is still formed by
unconnected tags. In some cases (i.e. [
We have presented the SOBOLEO system that makes it easy with its completely web based interface to jointly create a structured vocabulary and an index of web resources that is organized with this vocabulary. It combines a collaborative real time editor for a subset of SKOS with an easy to use annotation interface and a semantic search in the annotated resources.
We believe that SOBOLEO is already a useful tool, however, it is not as powerfull as we would like. For the near future we plan: • Better evolution support for the editor, including undo and joining of concepts. • Better support for the social aspects of collaboration, like private annotations or multiple communities within one SOBOLEO instance.
This work was funded in part by the German Federal Ministry of Economic and Technology under the project KSIunderground and by the German Federal Ministry of Education and Research under project Im Wissensnetz.