Categories Programming

Continued from Introduction

To ground our expectations for what enterprise IT can gain from component-based software, let's consider the benefits that have been promised at different times as component technology has evolved.

At one point, proponents argued that component-based software would lower the barriers to entry into the mass-market software marketplace. If a word processor, for example, were structured as a collection of components, then a small development company could enter the market with specific enhancements to the document preparation process. An entrepreneurial team could focus on a specific area of expertise, without the need to duplicate the entire core capability of a product such as Microsoft Word.

With this model of component software in mind, visionaries suggested that word processors, spreadsheets, and other end user applications would be extended in a manner determined by users' needs--rather than being tied to one or two large vendors' schedules for monolithic updates, such as the massive Word 97 (or the even more complex forthcoming Word 2000).

In practice, though, the potential for a flatter and more competitive market in general-purpose software was neutralized by other developments. Individual products such as word processors and spreadsheets were absorbed into multifunction suites, with extensive coupling between different applications.

While it was true that a single module might provide spell-checking capability for a word processor, a spreadsheet, and a presentation graphics tool, the interfaces that made this happen were obscure. Standard software worked well enough to discourage most users from exploring third-party enhancements, which might create compatibility problems. New file formats propelled users forward on a continuing treadmill of upgrades to core applications, rather than letting them add new functions to a stable productivity platform.

Productivity gains, if any, accrued to the vendors of mass-market software rather than to the buyers. Even for vendors, however, it is far from certain that component techniques created significant opportunities for software reuse: Commercial PC applications pursued an evolutionary path that was twisted and kinked by competitive pressures, such as the sudden burst of demand for Web-enabled applications.

The first wave of components, therefore, delivered only a jigsaw-puzzle kind of modularity. Yes, applications could now be defined in terms of pieces with specific interface conventions, but the pieces might fit together only in one useful way. When it came time to write a major enhancement of an application, all of the pieces needed to be overhauled together. The result might be a more manageable development process, but it's not clear that costs were reduced or that software quality was enhanced.

Next: What Happened Next?

Published as Enterprise Computing in the 4/20/99 issue of PC Magazine.