Name
Abstract | ||
|---|---|---|
Modularity of feature representations has been a long standing goal of feature-oriented software development. While some researchers regard feature modules and corresponding composition mechanisms as a modular solution, other researchers have challenged the notion of feature modularity and pointed out that most feature-oriented implementation mechanisms lack proper interfaces and support neither modular type checking nor separate compilation. We step back and reflect on the feature-modularity discussion. We distinguish two notions of modularity, cohesion without interfaces and information hiding with interfaces, and point out the different expectations that, we believe, are the root of many heated discussions. We discuss whether feature interfaces should be desired and weigh their potential benefits and costs, specifically regarding crosscutting, granularity, feature interactions, and the distinction between closed-world and open-world reasoning. Because existing evidence for and against feature modularity and feature interfaces is shaky and inconclusive, more research is needed, for which we outline possible directions. |
Year | DOI | Venue |
|---|---|---|
2011 | 10.1145/2019136.2019142 | SPLC Workshops |
Keywords | Field | DocType |
feature interaction,feature-oriented software development,modular solution,corresponding composition mechanism,modular type checking,feature representation,feature-oriented implementation mechanism,feature module,feature modularity,feature interface,granularity,interfaces,information hiding,software development,modularity,variability | Cohesion (chemistry),Type checking,Computer science,Information hiding,Theoretical computer science,Modular design,Granularity,Software development,Modularity | Conference |
Citations | PageRank | References |
26 | 0.80 | 53 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Christian Kästner | 1 | 3591 | 135.92 |
| Sven Apel | 2 | 3980 | 184.13 |
| Klaus Ostermann | 3 | 457 | 23.97 |