Future Research Directions
We are at present extending our work in several directions.
First, we are reﬁning the heuristics of our tool so as to make them more accurate.
Second, we are planning to enforce our ﬁndings by using other legacy systems. As a starting point, we will use an RM-COBOL system that comprises 319 programs and 2018 2 8 lines of code.
Third,we will investigate whether our ideas hold for legacy systems written in other languages, such as in C or Java.
Lastly, even when we found that using direct migration certainly has incidence in service quality at the service frontier level, intuitively many WSDL anti-patterns might be actually introduced by the tools employed to build the corresponding WSDL documents.
Indeed, recently, we have shown that there is a statistical correlation between the WSDL anti-patterns present in a SOA frontier depending on the code-ﬁrst WSDL generation tool being employed.
We are therefore planning to extend this analysis to other languages and therefore WSDL construction tools.
Our utmost aim is to determine how much WSDL anti-patterns occurrences are explained by the approach to migration itself, and how much of them depend on the WSDL tools used.
Another research direction is taking into account other design service quality metrics for improving the results of applying the assisted migration. These metrics might be used for assisting developers to make decisions when they are performing the manual part of the assisted migration. For instance, these metrics can be used to identify real services from a list of candidate services.