1 / 15

Designing Software for Ease of Extension and Contraction

CSE870 Miniproject Presentation. Designing Software for Ease of Extension and Contraction. Ali Ebnenasir & Farshad A. Samimi {ebnenasi,farshad}@cse.msu.edu. Main Paper. Problem: How to extend and change software after its initial design? Obstacles: Excessive information distribution

dawn
Télécharger la présentation

Designing Software for Ease of Extension and Contraction

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CSE870 Miniproject Presentation Designing Software for Ease of Extension and Contraction Ali Ebnenasir & Farshad A. Samimi {ebnenasi,farshad}@cse.msu.edu

  2. Main Paper • Problem: How to extend and change software after its initial design? • Obstacles: • Excessive information distribution • Chain of data transforming components • Integration of many functions in one • Dependency loop (nothing works until everything works) • Answer: Design for change (extensibility and removability)

  3. Main Paper (cont’d) • How to design for change? • Requirement identification (minimal subset of family members) • Information hiding (modules and their secrets) • Virtual machine design • Uses relation hierarchy • Example: • An address processing system (A set of programs to read in, store, and write out lists of addresses) Main Algorithm Input Processing Output Processing

  4. Main Paper (cont’d) • What if the format of the input address is changed? • Design another program from scratch • Design for change (A hierarchy of virtual instructions) add INADFO(frmttype frmt) INADFO: Reads in an address in a format, specified by a parameter addData INADSEL(int frmt) addfrmttype INFSL(frmttable tab,int frmt) INADSEL: Reads in an address using one of a set of formats INAD(datablock, int frmt, dev strg) INFSL: Selects a format from a format table INADFO INAD: Reads in an address and store it. The address is assumed to be in a specified format INFCR INADSEL INAD INFSL INTABEXT INTABCHG INFDEL

  5. First Related Paper: Commonality Analysis • Problem: How to identify family members? • Answer: A systematic approach in the analysis phase • Defining family members: • Sources of abstraction: Terminology, Commonalities • Variations of a family member • Variability: The scope of the variations of the entire family • Commonality Analysis (CA) Elements: • Terminology, Commonalities, and variability • Output: Documented family requirement information • User: Designers [David M. Weiss, 1997]

  6. First Related Paper (cont’d) • CA Process: Based on Regular Meetings • Steps (sequential) • Preparing: the required sources for initial sessions • Planning: the goals and the scope of the analysis • Analyzing: characterization of family members • Quantification: defining the parameters of the variations • Reviewing: by an external reviewer • CA is a part of FAST at Lucent Technology (since 1992) • Motivated by Parnas’ idea of design-for-change [David M. Weiss, 1997]

  7. Second Related Paper:A Model for Designing Adaptable Software Components • Problem: How to design adaptable components? • Design-for-change (Parnas idea) • Efficiency (component’s behavior) • Application builders • Language independency (implementation) • Legacy codes • Major challenges: • Providing environmental information for the components Proposed approach  Using Arbitrator • Designing interfaces Proposed approach  Active Interface [George T. Heineman, 1997]

  8. Second Related Paper (cont’d) • Implementation Strategy • Framework (ADAPT specification language) • Example: • adding new functions to a Spreadsheet • Advantages of language-base approach: • From component designer’s view • From applications builder’s view • Efficiency of components • Re-engineering of legacy codes [George T. Heineman, 1997]

  9. Third Related Paper:Software Component Technologies and Space Applications • Problem: Traditional view of one-of-kind products • Answer: Component technologies (domain-specific design-for-change) • Challenges: • Encapsulation • Composition of components • Paradigm • Scalability • Verification • Key issue: Paradigm shift [Don Batory, 1995]

  10. Third Related Paper (cont’d) • Implementation proposal: Design maintenance • Domain modeling (program families, requirement identification?) • Modifications required in other parts • Reflective computations • Extending Parnas’ idea: • Generating efficient code automatically • Motivation in space domain:productivity [Don Batory, 1995]

  11. Indirectly Related Paper 1: Refinement and Separation of Concerns • Problem: Design-for-change • Basic idea: Any changes should be reflected in all design aspects • Solution: Change the definition of module • Module: • Basic unit of abstraction • Encapsulates source code, documentation, formal properties, and performance model [Don Batory, 2000]

  12. Indirectly Related Paper 1: (cont’d) • Refinement: Abstraction of a common feature • Family of similar applications based on the refinement • Design methodologies for architecturally changeable and extensible systems (e.g., GenVoca) • Analyzing the effect of a change on all design aspects [Don Batory, 2000]

  13. Indirectly Related Paper 2:Issues in the Design of an Extensible Operating System • Problem: Overcoming obstacles in designing (dynamically) extensible operating systems • Main observation: • Requires extensible software infrastructure Design-for-change  Parnas principle! • Domain-specific goals: • Incrementality • Correctness and integrity • Efficiency [S. Savage and B. Bershad, 1994]

  14. Indirectly Related Paper 2 (cont’d) • Dynamic extensions and extensibility namespaces • Parnas’ VM • Parnas’ Uses structure • OS domain-specific major concern: Correctness • Conflicting resources • Verification • Protection ( by design ) • Automating protection [S. Savage and B. Bershad, 1994]

  15. Conclusion • The outcome of the main paper: • Design-for-change principle • Methodology • The impact of the main paper • Analysis, Design, and Maintenance phases • Design-for-change affects reusability and adaptability

More Related