Alexander L. Wolf
University of Colorado at Boulder
alw@cs.colorado.edu
http://www.cs.colorado.edu/users/alw/
Abstract
Most of the energy in software architecture research during the past ten years has gone toward creating formal notations, design environments, and assessment practices for software architects. This activity has gone on largely in parallel with, and independent from, the effort that has gone into the development of middleware component frameworks such as CORBA and Enterprise Java Beans. Has the energy put into software architecture research paid off?
Unfortunately, there is currently a higher cost to developing, maintaining, and analyzing architectures than there is benefit to having them. Compare this to what proponents of component frameworks have to offer. Generally speaking, they provide a ready-made solution to the challenges of modern, complex software development: Their frameworks inherently support distributed systems, they are compatible with popular object-oriented design and implementation techniques, and they are supported by commercial products. It would be hard to argue that support for software architecture comes close.
The counter argument is that component frameworks support only a relatively specific architecture, while the field of software architecture is trying to address more general structures. This is quite reasonable. But if that is the case, then software architecture technology should in some way be useful to the developers and users of component frameworks. A reasonable challenge for the software architecture research community, therefore, is to demonstrate the ability for software architecture technology to support or enhance middleware component framework technology.
Last Updated: May 22, 2000 by Elisabetta Ferrando