Reuse Success - When and How?

Ernesto Guerrieri

Inso Corporation

299 Promenade Street, Providence, RI 02908

Tel: (401)752-4464, fax: (401)752-4444

Email: eguerrieri@inso.com, {e.guerrieri@ieee.org, guerrieri@acm.org, guerrier@bu.edu}

URL: http://www.inso.com/

During the panel "The Future of Systematic Software Reuse" at ICSR-3, Rubén Prieto-Díaz predicted that, by the year 2000, "software reuse will cease to exist" (not in the sense that it will die, but in the sense that it will be part of the mainstream software engineering process) and that "the ultimate success of software reuse will be marked by its disappearance" [Prieto-Díaz94]. Others have stated that we have solved the reuse problems and some complain that we are seeing repeat solutions to problems already solved. So why does software reuse exist? What problems are left to be solved? Have some of the problems changed? Where do we go from here?

More fundamental, if we believe that software reuse will succeed, when will we know that it has succeeded and what needs to happen in order for it to succeed?

First, I do not think we have yet succeeded in software reuse. In the State of the Practice (SOP), I do not see alot being applied. Yes, OO class frameworks, Java, Visual Basic, etc. have facilitated the code reuse, but less is done, in general, at the higher level such as in the design and requirements phases of a project.  In the State of the Art (SOA), I have not perceived any incremental progress although many issues have been addressed. Second, the accomplishments of software reuse are partly due to the software reuse community, but mainly external activities are helping achieve it. Third, as a community we should be able to define a roadmap to the success of software reuse and WISR should take on the role of reviewing the progress to this achievement.

We need to understand where we stand with software reuse. Did we make any noticeable progress this year. Although there may be individual projects that succeed at any given point, what impacts the whole community? Which ones are within the reuse community? Which derive from outside the reuse community and impact us? What are the research projects that need to be addressed by the SOA? Where are the roadblocks with the SOP and what can help eliminate them? These are the issues that the WISR community should be addressing and providing guidance.

The reuse community has provided many key solutions on many of the reuse issues. Just to name a few, there has been alot of concentration around reuse libraries, metrics, domain analysis, component-based development, management issues, product lines, domain specific software architecture (DSSA), and generators. We also see alot of reuse books that have been published. Alot of the concepts promoted by software reuse are also considered good software engineering practices. This sounds great, but most of these areas are not of widespread usage.

How did the success of the following products impact the SOA/SOP of software reuse:

• Visual Basic?
• Power Builder?
• Access?
• SAP R3?
• Java?
• WWW Browsers?

How have the following technologies impact the SOA/SOP of software reuse:

• OO Languages?
• Patterns?
• Frameworks?
• COM/DCOM?
• CORBA?
• Parameterized programming?
• Client/Server Architectures?
• Use Cases?
• UML?
• WWW?

Overall, the successes of products and technologies outside the reuse community have had more impact on the SOP of software reuse than any work from within the reuse community. How can we foresee this in the future and provide the technology transfer of the SOA from the reuse community to the external successes? ICSR has provided the reverse technology transfer by presenting many of these technologies as keynotes. Various individuals in the reuse community have attempted to evangelize software reuse within the communities of these other technologies, but with mixed results (one reason could be that they were outnumbered by short-sighted issues within these other technologies).

What I would like to see is a more leadership role by the WISR reuse community to:

• define what is meant by the success of software reuse,
• review the progress in the SOA/SOP in software reuse against some metric,
• set the goals for future progress in the SOA/SOP in software reuse,
• provide a roadmap to achieve these goals, and
• facilitate the technology transfer to a wider audience/population.

We should all leave the workshop with clear objectives of the future activities in software reuse research, implementation, and technology transfer. One should attend future workshops with the intent to analyze the progress made against the goals, to understand the impact of other technologies and products, and to correct/update the WISR Roadmap to Reuse Success.

There have been various papers in the literature (including past WISR proceedings) that have addressed the adoption, leverage, and institulization of software reuse [Collins95, Czarnecki97, Diaz-Herrera95, Frakes94, Gaska95, Griss95, Johnsson95, Kelly95, Lam97, Moore97, Neighbor94, Steyaert97, Price97].

There have been various Reuse Capability Models proposed in the past that would help capture the state of reuse in an organization (but not in the general practice). Most of these were based on the SEI's Software Capability Model. Some reuse maturity models were defined by organizations and projects such as the REBOOT project, Harris Corp., and SPC.