Information Technology Reference
In-Depth Information
Identifying existing solutions. Sure Step does not support the automatic identifica-
tion of existing solutions. The new approach should be able to identify services based
on use cases and linked requirements. More specifically, a novel integrated approach
should propose a list of candidate services and provide service descriptions.
Narrow down service discovery results. In our discussions consultants empathized
that the time to discuss solutions with customers is limited and that they would need a
feature supporting them by highlighting the most promising solution(s). A novel
approach should highlight the fact that this service only has a low priority and that it
is unlikely that it can contribute to the envisioned solution.
Provide details on services of interest. Details of a service of interest need to be
immediately accessible for further discussion. The new approach should allow con-
sultants to access important information related to a selected service.
3.2 Identifying Research Informing S 3 C
An analysis of existing work shows that similarity analysis is a well-described issue in
ERP literature [8, 9]. However, we could not identify relevant work focusing on
service discovery mechanisms for similarity matching. Also in requirements research,
little has been reported on service discovery. Schmid et al. [10] discuss a require-
ments-led process enabling runtime service discovery but do not report on tool sup-
port. Elsewhere, Esmaeilsabzali et al. [11] present new models for requirements-based
service discovery that assume formal expression of system operations. Zachos et al.
[12] have researched new tools and techniques to form service queries from incom-
plete requirements specifications as part of the EU-funded SeCSE Integrated Project
[13]. We considered the work by Zachos et al. [14] to be most promising for extend-
ing Sure Step.
The SeCSE requirements process is depicted in Figure 3. Service queries are ex-
tracted from a service request constructed from a requirements specification and then
fired at service registries. The retrieved service descriptions enable consultants and
customers to select the most appropriate service(s). The main innovation is to expand
service queries to handle requirements expressed in natural language. As such, SeCSE
appears to be the first approach to integrate requirements and service discovery meth-
ods and tools. Therefore it was chosen to be the base for the S 3 C project.
Fig. 3. SeCSE's Requirements Process
Search WWH ::




Custom Search