International Journal of Web Services Practices, Vol. 3, No.3-4 (2008), pp. 129-135
ISSN 1738-6535 © Web Services Research Foundation, Seoul, Korea
129
Abstract— Rapid growth of published Web services makes
their discovery more and more difficult. Many approaches are
proposed to solve this discovery issue. Most of them are interested
only in Web service’s description itself and neglect the
user-centric aspect of the discovery process. User’s involvement
may be seen as contributing information, after using or testing a
service, which marks the personal opinion on service’s
functionality, quality or invocation cost. This kind of contribution
is perceived as a participative aspect inside the Web 2.0
environment and can enhance the Web service discovery process.
We illustrate in this research work, a collaborative tagging-based
environment for Web service discovery, allowing users to tag or
annotate a Web service using keyword or free-text. Our system
proposes consequently two types of query to search tagged Web
services: keyword-based and free-text. We put in place an
advanced mode in the discovery by keyword which offers
different ways to combine keywords together inside a query to
make it more flexible and accurate. In both types of look-up (by
keyword and by free-text), synonym relation between terms is
processed in order to improve the similarity computing between a
query and the tagged web services. The Web services found after
a discovery process are ranked according to their tags’ weight, in
the case of discovery by keywords.
Index Terms— Web Service Discovery, Collaborative Tagging
System, Vector Space Model, WordNet
I. INTRODUCTION
ROWNING into a great number of publicly accessible
Web services, a user (a neophyte Web user who needs to
start with Service Oriented Computing, a designer or developer
of Service Oriented Applications) can be confronted with a
difficult exercise when trying to find a Web service relevant to
a requirement. Generally, these users have difficulties in
Manuscript received 5
th
December 2008.
Uddam CHUKMOL is a doctoral student of INSA Lyon and he is
conducting his research activity within the SOC (Service Oriented Computing)
team of LIRIS laboratory – Université de Lyon, INSA-Lyon, LIRIS, UMR5205,
F-69621, FRANCE (e-mail: uddam.chukmol@insa-lyon.fr).
Aïcha – Nabila BENHARKAT is an Associate Professor in the Information
Technology and Computer Engineering Department of INSA Lyon and she is
also a researcher in the SOC (Service Oriented Computing) team of LIRIS
laboratory – Université de Lyon, CNRS, INSA-Lyon, LIRIS, UMR5205,
F-69621, FRANCE (e-mail: nabila.benharkat@insa-lyon.fr).
Youssef AMGHAR is a Professor in the Information Technology and
Computer Engineering Department of INSA Lyon and he is also a researcher in
the SOC (Service Oriented Computing) team of LIRIS laboratory – Université
de Lyon, CNRS, INSA-Lyon, LIRIS, UMR5205, F-69621, FRANCE (e-mail:
youssef.amghar@insa-lyon.fr).
exploring Web services in a few steps and, to the best that we
know, nothing is available yet to facilitate the construction of
their requests in a free-style, flexible, intuitive but efficient way.
Most of the time, users discover Web services through a search
interface using text based query with some advanced features
or they have to construct their queries in a complicated
formalism to express their searches’ requirement. We argue
that both styles of query do not facilitate the Web service
discovery. The reasons behind this affirmation are: firstly,
simple keyword based discovery mechanism frequently offers
low matching result in the current environments because they
deal with poorly described Web services (insufficient textual
description of Web services within WSDL files). Secondly,
complicated formalisms used to construct discovery’s queries
are often description logic based, which require efforts from
not only the users while constructing their queries but also from
the service’s providers to enrich their published services with
description logic or ontological information. The search
algorithms, in this case, have recourse to ontology usage and
demand standardized ontology definition.
Besides, we can consider the Web as the simplest
environment for neophyte users to explore Web services as if
they were executing an information lookup process. In effect,
after the official shutdown of UDDI project in 2006, Internet
has gained a real popularity as a global repository of Web
services. We are able to find different Web sites providing
access to Web services such as XMethods [23],
WebServiceX.NET [25], StrikeIron [24], etc. Different
existing works emphasize and confirm this statement. In [26],
Internet is perceived as a natural source of Web services
because Web service’s descriptions are hosted in various Web
servers accessible by Web crawlers. A collection of Web
services crawled from the Web is provided in [22]. In a simpler
way, [28] uses different general purpose search engines in the
service discovery process. These works allow us to argue that
there are several interesting access points to active Web
services that can be found on the Internet and using this world
wide environment will bring Web services a large step closer to
mass public users.
Despite the fact that there are no sufficiently rich hyperlink
structures between WSDL documents (Web service’s
description), a Web based crawler can still be operational to
collect this kind of files from the Internet [27]. Building a
collection of Web services that are extracted from the Web can
assure not only the public access to the services but also the
regular update of services. We, henceforth, will focus our work
Enhancing Web Service Discovery by using
Collaborative Tagging System
Uddam CHUKMOL, Aïcha-Nabila BENHARKAT and Youssef AMGHAR
D