Cross-Family Architecture as a Reuse Mechanism

Even though a survey of 50 large organizations by Cutter Consortium found that nearly half (48 percent) are committed to component reuse, only 37 percent have a specific group designated to facilitate it. "That difference reflects those who are committed in an abstract way and those that have a real day-to-day commitment," states Cutter Consortium Senior Consultant Paul Harmon.

Given those numbers, the subject of component reuse and, more specifically, the cross-family architecture reuse mechanism were recently discussed at a meeting of the Cutter Technology Council. A cross-family architecture supplies a common infrastructure for all products within a domain. This allows an increased capability to deliver software products swiftly, respond to short-term needs, and achieve improved quality.

Cutter Technology Council Fellow Tom DeMarco, in the latest Council Opinion, asserts that a willingness to invest in a cross-family architecture reuse program will distinguish successful from unsuccessful IT user organizations. "The superiority of a cross-family architecture is that the similarities are stronger within a local domain than across the market. This is both its strength (commonalities tend to be stronger) and its weakness (you are going to have to build much of your own infrastructure)."

A downfall is cost. Companies must be willing to view cross-family architecture as an investment with the returns growing in time. For those organizations, the Council majority recommends the following actions:

* Establish a policy for investment in system-building infrastructure.

* Perform a census of domains served by the organization.

* Pick the most important domain and apply the investment budget to building a cross-family architecture for that domain.

Although it is important, Cutter Technology Council Fellow Ed Yourdon does not see reuse as an all-or-nothing phenomenon. "I see no compelling argument that the cross-family architecture strategy will succeed when so many other approaches to reuse are alleged to have failed. Different organizations achieve different levels of reuse, depending on how important they think it is. Even the oblivious organizations achieve about 10% reuse -- without measuring it, without providing rewards for it, and without really understanding how much benefit they get from it."

On whether the use of a cross-family architecture will determine a company's success or failure, the Council remains divided.