Understanding feature modularity
Year: 2006
Type: dissertation
Abstract: Features are increments in program functionality. Feature abstraction, the process of \nabstracting programs into their constituent features, is a relatively common yet informal \npractice in software design. It is common because it simplifies program understanding. It \nis also important for software product lines whose essence is the systematic and efficient \ncreation of software products from a shared set of assets or features, where each product \nexhibits common functionality with other products but also has unique functionalities. \nThus, it seems natural to modularize feature abstractions and use such modules as \nbuilding blocks of programs and product lines. Unfortunately, conventional modularization \napproaches such as methods, classes and packages are not geared for supporting feature \nmodules. They present two recurrent problems. First, a typical feature implementation \nis spread over several conventional modules. Second, features are usually more than \nsource code artifacts as they can modularize many different program representations \n(makefiles, documentation, performance models). \nAn undesirable consequence is that developers must lower their abstractions from \nfeatures to those provided by the underlying implementation languages, a process that is \nfar from simple let alone amenable to significant automation. The conceptual gap created \nbetween feature abstractions and their modularization hinders program understanding and \nproduct line development. The root of the problem is the fact that feature modularity is not \nwell understood and thus not well supported in conventional programming languages, \nmodularization mechanisms, and design techniques. \nIn this dissertation, we explore language and modularity support for features \nfounded on an algebraic model geared for program synthesis. Our model integrates ideas \nfrom collaboration-based designs, mixin layers, aspect oriented programming, multidimensional \nseparation of concerns, and generative programming. We assess our model \nwith an implementation of a non-trivial product line case study, and evaluate feature support \nin emerging modularization technologies.
Authors: Don Batory, Roberto Erick Lopez Herrejon
Institution: The University of Texas at Austin
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1774095927
