Design For Reuse

(click to go to roadmap)

WISR 4 - Design For Reuse (Frakes)

Goal: Explore ways of creating reusable lifecycle objects (requirements, designs, code, test cases, etc.)
Approach: bottom up, obtaining design principles from lower level LCOs (Assumption: we know many lower level priniciples)
Results:
- Three Programming Languages: Ada, C++, C
- Defined a number of lower level principles, but disagreed about scaleup
- 3Cs - Top down approach for deriving reuse design - had problems finding adequate design principles

WISR 5 - Design For Reuse (Weide, Lea, and Frakes)
- Discussed guidelines for reuse - general, Ada, and/or C++
- Issues::
  - How components measure up
  - How do design oriented features affect reuse
  - Need for extra-linguistic tools, specifications, and design methods
  - Need for reuse metrics
  - Scalability
  - Practically useful principles for components (and, oh yes, other design entities, whatever they may be)
  - Handbooks will be huge
  - Teaching must be "operationalized"
  - Basic principles exist: black-box components, encapsulation, unbreakable abstraction, and extension by addition only (hmmm... local certifiability)
- WISR 6: Design For Reuse and OO Methods (Lea and Frakes) (click to go to report)
  - Goals: Language Independent principles (based on WISRs 4, 5, and 6)
  - Issues:
    - Common realizations as they apply to (C++, C, and Ada)
    - Rationales
    - Process by which agreement on principles reached
  - Major Principles:
    - Component Structure
    - Interfaces
    - Composition
    - Parameterization
- WISR6: Language Issues for Generic Code Architectures (click to go to report)
  - Initial Goals: To determine sets of features and mechanisms useful for constructing systems from generic architectures
  - Revised Goals:
    - Exploration of architecture commonality through a transformational model of program generation
    - Definition of standard vocabulary, concepts, and mechanisms wrt above mechanism
    - attempt to map relations between generic architecture systems via the standard.
  - Systems:
    - Synapse
    - Predator
    - RESOLVE
  - Observations
- WISR 6 - Hybrid Reuse - Domain Specific Kits (Griss) (click to go to overview)
- WISR 7 Micro-Architecture of Software Components and The Need For Good Mental Models of Software Subsystems (click to go to report)
  - Micro-Architecture Goals: to focus on
    - the structure and behavior of software component interfaces
    - the implementations of individual components, sets of components, and
    - how components compose and "interoperate" with each other
  - Mental Modeler Goals: to focus on how software engineers form internal mental models of the things they interact with in order to understand those interactions.
- WISR 7 Object-Oriented Reuse - A Tale of Two Cultures (click to go to report)
  - Goal: To explore the apparent disconnect between OO community and systematic reuse community. We believe that each community has something to offer the other.
  - Reasons for Disconnect:
    - ``Hacker/builder'' vs ``methods/software engineer'' culture
    - OO community works in small teams, largely oral culture.
    - Reuse community's evidence for successes not crisp or compelling enough for OO community.
    - Systematic reuse methods not operational, mechnical or concrete enough for use/adoption by OO community.
    - Current OO practice can get by without much reuse community advice and knowledge
    - larger scale problems, systems, libraries or teams need a more structured approach.