| dc.description.abstract | Representing data outside of and between programs is important in software that stores, shares, and manipulates information. Formats for representing information, varying from human-readable verbose (XML) to light-weight, concise (JSON), and non-human-readable formats (TLV) have been developed and used by applications based on their data and communication requirements. Writing correct programs that produce information represented in these formats is a difficult and time-consuming task, as developers must write repetitive, tedious code to map loosely-typed serialized data to strongly-typed program objects. We developed S.IM.PL Serialization, a cross-language multi-format information binding framework to relieve developers from the burdens associated with the serialization of strongly-typed data structures. We developed type system scopes, a means of encapsulating data types and binding semantics as a cross-language abstract semantics graph. In comparison to representing data binding semantics and information structure through external forms such as schemas, configuration files, and interface description languages, type system scopes can be automatically generated from declarations in a data binding annotation language, facilitating software engineering. Validation is based on use in research applications, a study of how computer science graduate students use the software to develop applications, and performance benchmarks. As a case study, we also examine the cross-language development of a Team Coordination (TeC) game. | en |
