There is now a considerable body of work arguing for the importance of industrial clusters, linkage, and spatial agglomeration to innovation and economic dynamism. Increasingly, as well, this work has been drawn on in policy terms by those seeking to emulate the success of such iconic high-technology clusters as Silicon Valley or Boston's Route 128. In this paper I first provide a brief review of this body of work, emphasising the different theoretical and conceptual traditions on which it has drawn and identifying shifts in emphasis over time. I then present a critical appraisal of the attempt of the State Government of Victoria, Australia, in the late 1980s and early 1990s to boost innovation and technologically advanced economic development by means of spatially focused technology precincts. I look at the origins of the initiative itself, including the role of international consultants in promoting knowledge transfer in the policy arena and a particular model of technology-based economic development. I then present an analysis of the implementation of technology-precinct strategy in practice, and an account of the legacy of what is, at one level, a crude and indeed largely unsuccessful attempt to emulate high-technology growth based on a model of agglomeration, linkage, and technology transfer. I conclude by noting that contingent factors at state and national levels, including economic shifts and political change, account, in part at least, for the fate of technology precincts on the ground. There are more fundamental problems, however, inherent in attempting to base strategies for economic development on the sort of high-profile international exemplars around which the now considerable body of work on industrial districts and knowledge transfer, innovation, and technology development has developed. New industrial districts, innovation, and the learning region Initial accounts of new industrial districts, drawing on a related set of ideas based around post-Fordism and flexible production, tended to see clusters in neo-Marshallian terms (