Detecting and predicting amorphous and ambiguous anti-patterns of RESTful URI schema toward improving its discovery

REpresentational State Transfer (RESTful) technology exposes services as resource invocations, which are identified by Uniform Resource Identifiers (URI). The URI forms RESTful indexing schemas that match resources to their provided services, which makes the resources addressable and have a simple map format. Since the URI schema is the primary resource of RESTful service information, it should be readable by the service requesters. Unfortunately, many RESTful suffer from linguistic antipattern problems such as Amorphous (AMURI) and Ambiguous (AWS) anti-patterns, which make RESTful URI schemas difficult to read and understand. Therefore, their opportunity of being discovered is reduced. This research improved the readability of URI schemas to increase their discovery by devising two techniques: (1) a detection technique that detects AMURI and AWS anti-patterns in the original URI schema and (2) a prediction technique that suggests suitable patterns to replace the detected antipatterns in the rebuilt URI schema. This study used the design science research process to accomplish its objectives. Two expert review evaluations were used to validate the accuracy of the techniques. The first expert review manually detected such antipatterns from sample sets of original URI schemas, while the second detected such anti-patterns from rebuilt sample sets. The findings suggested an acceptable level of accuracy for both techniques, with 87.86% of AMURI and AWS anti-patterns detection and 98.1% of tidy and representative patterns prediction. Six readability metrics were used to evaluate the performance of both techniques by comparing the metrics values for the original and rebuilt URI schemas. The results demonstrate that the techniques improved the readability of URI schemas by 84.25%. Hence, the proposed techniques could improve the corresponding RESTful discovery.