With the increasing number of web services and growingr demand for web service oriented computing, it is necessary to develop fast algorithms that could take a service requirement, and through compositional search of UDDI registered web services, find a cost effective (in terms of length of service response, total cost of composed services, minimal length of services, amount of space used for composition discovery) result. This paper describes two approaches for syntactic search, one for discovery of web services, without chaining, and the second one for the compositional chaining of web services with a front-to- end DAG based search approach. The assumption is that a user provides a X m file stating his required output and set of inputs, and the algorithm is supposed to identify one or a chain of services that does the job.