ActiveRaUL: A Web form-based User Interface to
create and maintain RDF data
Anila Sahar Butt1,2 , Armin Haller1 , Shepherd Liu1 , and Lexing Xie2
1
CSIRO ICT Centre, firstname.lastname@csiro.au
2
Australian National University, firstname.lastname@anu.edu.au
Abstract. With the advent of Linked Data the amount of automatically
generated machine-readable data on the Web, often obtained by means of
mapping relational data to RDF, has risen significantly. However, manually
created, quality-assured and crowd-sourced data based on ontologies is not
available in the quantities that would realise the full potential of the semantic
Web. One of the barriers for semantic Web novices to create machine-readable
data, is the lack of easy-to-use Web publishing tools that separate the schema
modelling from the data creation. In this demonstration we present ActiveR-
aUL, a Web service that supports the automatic generation of Web form-
based user interfaces from any input ontology. The resulting Web forms are
unique in supporting users, inexperienced in semantic Web technologies, to
create and maintain RDF data modelled according to an ontology. We re-
port on a use case based on the Sensor Network Ontology that supports the
viability of our approach.
1 Introduction
The tools of choice for creating quality-assured ontology instances (the so-called
ABox ) are still ontology editors such as WebProtégé [4]. However, creating the ABox
in an ontology editor requires some degree of understanding of RDF(s) and OWL
since the user has to define to which class an individual belongs to and what are
the permissible relationships between individuals. To address this issue, some Web
publishing tools on top of Wikis, Microblogs or Content Management systems have
been developed (e.g. the work discussed in [1], [5] and [3]) that allow a user to exclu-
sively create ontology instances. However, they are mostly developed for a specific
domain (i.e. specific ontologies) and often do not strictly follow OWL semantics and
consequently allow the creation of a logically inconsistent ABox.
In this demonstration we will showcase ActiveRaUL [6], a Web service that op-
erates on a model defined according to the RDFa User Interface Language (RaUL)3
that consists of two parts, [1.] a form model describing the structure of a Web form
with different types of form controls, such as Textboxes, Radiobuttons, Listboxes etc.,
and their associated operations (CREATE, READ, UPDATE or DELETE) and [2.] a data
model defining the structure of the exchanged data as RDF statements which are ref-
erenced from the form model via a data binding mechanism. The ActiveRaUL Web
service also provides functionality to automatically generate a Web form-based user
interface according to the RaUL ontology from arbitrary ontologies. We argue that
the resulting user interfaces are easier-to-use for a semantic Web novice to create
3
See http://purl.org/NET/raul#
Deployment System Output
ssn:deploymentProcessPart
Page ssn:hasOutput Mult
Grap
ssn:hasDeployment DU
ssn:hasSurvivalRange
L:h Tran
ss
as
Da Inve
n:d
tem
taV
ep
Dynamic alu Data
ys
Group Button e
lo
S
Group
ye
ed
xsd:String
dO
ch
tta
nP
Dynamic
Button
n:a
lat
Group
ss
fo
rm
Group
ontology individuals,
Deployment and result in more accurate ontology instances than creating
Survival Survival
Group
them through traditional ontology Platform tools. We validated this hypothesis in
engineering
Process Part Range Property
a user study comparing our system with a state-of-the-art ontologyssn:hasSurvivalProperty
modelling tool.
2 Demonstrating the ActiveRaUL Web service
hasSubsystem
deployedSystem
5 1 6 3
Deployment System Output
hasOutput
deploymentProcessPart
hasDeployment DU
L
de
hasSurvivalRange
:ha
sD
plo
a taV
y
alu
em
ed
e 2
On
st
xsd:String
Sy
Pla
ed
ch
tfo
ta
rm
at
Deployment 4
Platform Survival Survival
Process Part Range Property
hasSurvivalProperty
Single-length Property Inverse Property Graph cycle
Multi-length Property Datatype Property Transitive Property
Fig. 1. Sub-graph structure for the System class in the SSN ontology
To relieve a Web developer from manually defining a Web form model according
to the RaUL ontology, we have extended the ActiveRaUL service with a deployment
endpoint that upon invocation generates RaUL Web forms from an arbitrary user
submitted ontology. The biggest challenge in automatically creating such Web forms
IgnoredInAlgorithm
from an ontology is the mismatch between the graph nature of RDF and the tree
structure of a Web form. In the algorithm implemented in ActiveRaUL we distin-
guish six different types of sub-graphs occurring in ontologies and introduce decision
controls to map these sub-graphs to useable web forms. We will demonstrate these
different types of mapping on a use-case based on the the Semantic Sensor Network
(SSN) ontology that can be used to describe the capabilities of sensors, the measure-
ment processes used and the resultant observations. Figure 1 shows the “System”
class of the SSN ontology and its relations to other classes, whereas Figure 2 shows
a screenshot of a generated Web form by ActiveRaUL of the “System” class. The
numbers 1–6 in both figures indicate the six different types of sub-graphs we distin-
guish in the algorithm (see Figure 1) and how they are displayed in the Web form
(see Figure 2).
Fig. 2. Screenshot of ActiveRaUL generated Web form for the ssn:System class
3 Evaluation
We compared ActiveRaUL to the widely used state-of-the-art ontology editing tool,
WebProtégé. The demonstration deployment of ActiveRaUL set up for the user study
already pre-loading the SSN ontology is available at: http://www.activeraul.org/
demo/index.html From the university deployment example defined by the SSN
working group we extracted three test cases, each with a number of tasks. We asked
users to model these test cases in WebProtégé and ActiveRaUL. For evaluating the
two systems we considered three usability metrics, (1) the effectiveness of the sys-
tem in supporting the user to complete the task measured by the accuracy of the
resulting models; (2) the efficiency of the users in using the system measured by the
time they spent on completing a task and (3) a user’s subjective reactions using the
system measured by the widely-used System Usability Scale (SUS) [2]. In the follow-
ing we briefly outline the results of our user study. These results are based on the
performance and feedback of twelve participants: five of which, based on their self-
assessment, were categorised into the semantics experienced user group, and seven
categorised into the semantics inexperienced user group.
Table 1. Overall accuracy in completing test cases in WebProtégé and ActiveRaUL
WebProtégé ActiveRaUL
Exp. Users 82.05% 91.03%
Inexp. Users 76.92% 87.91%
All Users 82.05% 91.03%
Accuracy: Table 1 shows the overall accuracy over the three test cases which shows
that the participants performed better in ActiveRaUL, managing to create 91%
correct triples compared to 82% in WebProtégé. For ActiveRaUL, the accuracy of the
participants was already very high in the first test case, even though no participant
has ever used the system before. This confirms our hypothesis that a Web form-
based user interface is familiar enough to computer literate users to create RDF data
correctly, even if the participants are inexperienced in semantic Web technologies.
Table 2. Average times (mm:ss) to complete test cases in WebProtégé and ActiveRaUL
Test Case 1 Test Case 2 Test Case 3
WebProtégé ActiveRaUL WebProtégé ActiveRaUL WebProtégé ActiveRaUL
Exp. Users 3:46 1:50 5:13 1:50 10:24 4:01
Inexp. Users 4:05 2:17 5:23 1:29 11:26 4:08
All Users 3:57 2:05 5:19 1:38 11:00 4:05
Efficiency: Table 2 shows the average times participants required to complete a
test case. Both participant groups, inexperienced and experienced, were significantly
faster (between 27% and 56% faster) completing the test cases in ActiveRaUL com-
pared to WebProtégé.
Usability: After completion of the three test cases in both systems we asked the
participants to rate their subjective reactions on the usability of the systems on
a five-point Likert scale as required by the SUS methodology. SUS yields a single
number representing a composite measure of the usability of a system with scores
in the range from 0 to 100, 100 being the best score. Overall ActiveRaUL scored
72.1 out of 100 points compared with 32.5 for WebProtégé, indicating that the
participants found ActiveRaUL easier to use than WebProtégé for the creation of
ontology instances.
Concluding, our user study proved that ActiveRaUL is indeed easier, more effec-
tive and more efficient to use for the creation of RDF data than the state-of-the-art
ontology editing tool.
References
1. J. Baumeister, J. Reutelshoefer, F. Puppe. KnowWE: a Semantic Wiki for knowledge
engineering. Applied Intelligence, 35:323–344, 2011.
2. J. Brooke. SUS - A quick and dirty usability scale. In P. W. Jordan, B. Thomas, B. A.
Weermeester, A. L. McClelland, editors, Usability Evaluation in Industry. Taylor and
Francis, London, 1996.
3. S. Corlosquet, R. Delbru, T. Clark, A. Polleres, S. Decker. Produce and Consume Linked
Data with Drupal! In Proceedings of ISWC, pages 763–778, 2009.
4. T. Tudorache, C. Nyulas, N. F. Noy, M. A. Musen. WebProtégé: A Collaborative On-
tology Editor and Knowledge Acquisition Tool for the Web. Semantic Web 4(1), 2013.
5. A. Passant, J. G. Breslin, S. Decker. Open, distributed and semantic microblogging with
smob. In Proceedings of ICWE 2010, pages 494–497, 2010.
6. A. Haller, T. Groza, and F. Rosenberg. Interacting with Linked Data via Semantically
Annotated Widgets. In Proceedings of JIST, pages 300–317, 2011.