英文文献翻译，自己翻译的不知道是什么东西，完全不能理解，求帮忙

Within the broader question of how best to organize the development and mercialization of a technology (Teece 1986), an innovator may choose to “open” its technology by allowing outsiders to participate in its development and mercialization (Shapiro and Varian 1998). Such strategies are particularly amenable to systems (Katz and Shapiro 1994, Marschak 1962) made up of multiple ponents, which can be opened one ponent at a time.1 Whether or not to open a new technology is understood to be one of the most crucial decisions an innovator will face. Opening has the potential to build momentum behind a technology, but could leave its creator with little control or ability to appropriate value (Katz and Shapiro 1986, Kende 1998, Morris and Ferguson 1993, Schilling 2009, Shapiro and Varian 1998, West 2003). Case study evidence suggests that an important element of this tradeoff relates to the impact of opening on various aspects of innovation and technical progress, such as the improvement of individual ponents; the creation of extensions, addons, and upgrades; the elimination of bugs and errors; and quality and cost improvements (e.g., Baldwin and Clark 2000, Kuan 2002, Langlois 1992, von Burg 2001, von Hippel 2005).

In this paper, I present empirical evidence on how different approaches to opening a system influence the rate of innovation.2 I consider two broad and fundamentally different approaches to opening: granting outsiders access to the platform, thereby opening up markets for plementary innovation around the platform, and giving up some control over the platform itself. Part of the difficulty in distinguishing these approaches is that they often coincide in theory and practice. If the platform owner devolves all control over the platform, there is no longer a party within the system who can restrict access to outsiders wishing to develop the platform or other ponents; consequently, widespread entry can follow (see Katz and Shapiro 1986, 1994). For example, much of the development of public telemunications and interputer munications networks over the last century has occurred around key platform technologies whose definitions were managed by public standards authorities. Thus, firms around the world could develop not just the core munications network equipment but also testing and installation equipment, specialized billing and messaging equipment, and other sorts of plementary ponents andservices (e.g., consulting). Analogously, open source versions of the UNIX operating system platform such as Linux have enabled firms to freely enter the markets for plementary applications software, hard-ware, and services.

However, platform owners can and often do open up plementary ponents without giving up all control. It is enough that independent developers of plementary ponents be granted permission to use the platform, while ensuring that their ponents are (legally and technically) interoperable with the platform and with each other (see Boudreau and Hagiu 2009, Farrell 2007, Merges 2008, Parker and Van Alstyne 2008). For example, Apple tightly controls development of its iPhone operating system (and the closely linked iPhone, iPod, and iPad hardware), but allows thousands of outsiders to develop software applications, mercial media, and usergenerated content. This example illustrates the important distinction between giving up control over the platform and simply granting access to the platform in order to open up plementary development.

Indeed, it is possible to discern strands of research focused more or less on these distinct notions of devolving control and granting access. For example, a number of theoretical papers have considered how granting wide access to independent developers of interoperating, mix-and-matchable ponents can foster vibrant markets with diverse ideas and active experimentation (e.g., Baldwin and Clark 2000, Farrell et al. 1998, Farrell and Weiser 2003, von Hippel 2005). A quite distinct strand considers the ability of platform owners to stimulate innovation by relinquishing control over foundational platform technologies (e.g., Katz and Shapiro 1986, Farrell 2007, Farrell and Katz 2000, Shapiro and Varian 1998).

In this paper, I am able to empirically distinguish the independent effects of granting access to and devolving control of the platform by studying a data set on the development history of handheld puting systems (1990–2004). These were relatively simple systems, where the hardware played the role of plementary ponents and operating systems played the role of platforms. Operating system platform owners at times varied the degree of access granted to outsiders (by changing their licensing policies, and sometimes also lowering entry barriers by sharing hardware “reference designs”). Platform owners also varied the level of control over their platforms by sharing equity ownership, narrowing their vertical scope, and allowing outsiders to contribute to isolated parts of the operating system (such as the graphical user interface).

The econometric analysis relates the rate at which new handheld devices were released to alternative modes and degrees of openness, within a robust count data panel framework (Wooldridge 1999). (As detailed within the paper, several important considerations motivate this particular approach to modeling and measurement.) I find that granting access to independent hardware developers was associated with a dramatic increase in the rate at which new devices were developed up to a fivefold acceleration. The liberalness of the licensing approach explains most of this acceleration, with intermediate policies (i.e., somewhat restrictive licensing) resulting in the highest development rates. Lowering entry barriers by sharing hardware designs also had a positive impact on development, but this effect was considerably smaller. Platform owners who went a step further by giving up varying degrees of control over their operating systems also experienced accelerated development, but the total effect was an order of magnitude smaller than that of granting access. Together, the various measures of giving up platform control account for a roughly 20% increase in development rates. (The precise estimate depends on the model specification, and the effect is not statistically significant in all models.) These results are robust to system fixed effects and covariates, various specifications of time controls, alternative functional specifications of the mean and error, and regressions on subsamples. I also confirm that the preswitch trends of switchers and nonswitchers are similar, to provide assurance that meaningful econometric parisons were made.

Both the empirical results and qualitative arguments based on the history of the industry suggest that fundamentally different economic mechanisms were set in motion by the two approaches to opening. By granting access to independent developers of plementary ponents, platform owners drew on a diverse set of capabilities and concepts while intensifying petition among these outsiders. In cases where platform owners also lessened their control over the platform, the results are consistent with a shift in the “balance of power” from the bottleneck platform owner toward outside developers and corresponding accelerated development by outsiders. However, the small magnitude of the effect calls into question whether incremental variation in platform control (as seen in these data) can do much to activate this effect. Further research is required to more precisely determine the nature and workings of the economic mechanisms set into motion by opening in these various ways and degrees.

This paper is organized as follows. Section 2 reviews the literature on open strategies in systems and innovation. Section 3 focuses on the distinction between opening the platform and opening plementary development. Section 4 describes the empirical context, and §5 presents the data. Section 6 describes the empirical approach. Section 7 presents the model results and robustness tests. Section 8 discusses the patterns that were found in the analysis. Section 9 concludes.