The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Lessons from History
The federal government had important effects on the development of relational databases. The earliest days of this subfield suggest that government missions can create new markets for technology, providing incentives for innovation. The Census Bureau's need to conduct a decadal census supported the information processing industry before computers were created to automate it.
Later on, government funding hastened commercialization. An example is the case of System R and Ingres. The critical issue is not which one was more successful or influential in the long run, but rather, to paraphrase a System R team member, whether either project would have succeeded in the absence of the other. The academic interest legitimized System R within IBM, and Ingres was bootstrapped off IBM's commercial influence.8 Competitive pressure, combined with the legitimacy bestowed on the relational model by government funding and academic interest, finally convinced IBM to sell relational database products. Were it not for the government-funded effort at UC-Berkeley, such databases probably would have been commercialized anyway, but later—and time-to-market is, of course, a critical factor with new technology.
That same example shows that the commercial interests of firms such as IBM can impede the continued development and commercialization of technologies that compete with existing product lines. IBM and its customers had vested interests in the established IMS technology and resisted change until external events proved that relational databases could become viable commercial products.
This case history also suggests that the large numbers of researchers passing through university laboratories, their willingness to share data and code, and their publication imperatives make university researchers ideal sources of technology transfer to the broader technical community. Industrial laboratories, by comparison, rarely place significant technologies directly into the public domain and have lower rates of personnel turnover, although they often benefit from greater and more stable supplies of resources. Especially in the computing industry, employees may take ideas into the marketplace on their own, but industrial laboratories are likely to publish only information that concerns completed projects or is not deemed critical to the company's vital interests.
Academic research is important for other reasons as well. Because it can push the cutting edge of technology and produce results that may evolve into commercially viable products, existing commercial suppliers never have a lock on advanced technology and are forced to respond to the marketplace of ideas.
Finally, in pursuing new ideas and new areas of technology, aca-