Baldwin Mirroring Hypothesis Part I
A new working paper has been released by Professor Carliss Baldwin and Lyra Colfer. This paper "The Mirroring Hypothesis: Theory, Evidence and Exceptions" is something that we should have written about before, however time is an issue. What is interesting about this paper is that it is Professor Baldwin's co-author, Lyra Colfer's Doctoral Dissertation. Her Dissertation has been abbreviated into one paper that provides substantial evidence that the methods being used in the development of People, Ideas & Objects software applications. Are not just valid, but also valid for the oil and gas industry at large.
In Part I of our review of this paper, I want to introduce the hypothesis, discuss the authors definition of Open Collaboration and note how People, Ideas & Objects deviates from their definition. And lastly in this first part, talk at length about a concept called Actionable Transparency.
The authors define the mirroring hypothesis as;
The mirroring hypothesis asserts that the organizational patterns of a development project (e.g. communication links, geographic collocation, team and firm co-membership) will correspond to the technical patterns of dependency in the system under development. Thus the hypothesis predicts that developers with few or no organizational linkages will design independent system components, while developers with rich organizational linkages will co-design highly interdependent system components. (The hypothesis claims a correspondence between organizational structure and technical architecture, but allows causality to flow in either direction.) p. 1Open collaboration as it is called in this paper is defined as:
In an open collaborative development project, product design information, such as software source code, is placed in the public domain. Independent entities including individuals and firms contribute voluntarily to the design, according to their own private needs and interests; they self-select their contributions without relying on managers or market prices to guide them (Raymond 1998, 2001; Benkler 2002; von Hippel and von Krogh, 2003; Weber 2004; Lakhani and von Hippel, 2009). The open collaborative literature extrapolates from these observations, and uses the mirroring hypothesis to predict the structural form of products created in these settings. p. 12A clarification is necessary to reconcile People, Ideas & Objects differences to this definition. That is the source code is not placed in the public domain. The costs associated with development are too high to expect a volunteer group to be able to identify and build the necessary scope of an application of this size. Access to the code is made available, however, with the desired benefit that the direction the application takes is determined by the user groups based on their needs. I detailed the differences between People, Ideas & Objects and pure Open Source developments in a recent post.
This difference between the definition and our development do not preclude us from learning from this paper and applying its conclusion. On the contrary, I think that the difference between the theoretical and commercial world, in terms of the conclusion, is negligible, and its conclusion is precisely applicable.
6 Actionable Transparency
"Actionable Transparency" is a term that has been coined by Professor Carliss Baldwin. As we will see in Part II of our review, the authors determine that the situation that we fall within "break the mirror" and it is through the concept of Actionable Transparency that we find the results of this research interesting and applicable.
In The Age of the Smart Machine, Shoshana Zuboff (1988) observed that the increasingly information-based nature of industrial work has radically increased its “transparency.” She argued that, when auto-generated archives constitute a near-perfect surrogate for the activities that generated them, access to those archives provides “universal transparency” into what others are doing (pp. 315, 356-361). In effect, anyone with access to the archives can “see” what’s going on without the benefit of direct input or assistance from others. p. 28Google would certainly subscribe to this theory. If everything is discoverable, then we certainly have a different approach to how we do things.
Material transparency denotes the mere disclosure of information. By comparison, conceptual transparency requires not only that contributors can access the information, but also that they can make sense of it (cf. Wenger, 1990). Finally, actionable transparency requires not just that they can make sense of it, but also that they can act on it (cf. West and O’Mahony, 2008). p. 28We come to a problem in the development of the People, Ideas & Objects systems that the needs of one user can not necessarily be met through a generic design. I would assert that we are not seeking a generic design, but one that deals with the unique nature of the oil and gas industry as represented in the Joint Operating Committee. How this comes about is unknown, however, lets accept that we can have a multitude of different opportunities and options available to users of the system. If we make the assumption that it is possible, I think we can see in these authors work that Actionable Transparency is the solution.
Material and conceptual transparency do not imply actionability, however. Just because a potential designer understands a design doesn’t mean she can act on it. However, if she has the right and means to customize her own private copy of the design, then she has more room for action. The ideal form of actionable transparency goes even further: the designer can combine her changes with a host of others’, in near real-time, while at the same time guarding against design conflicts and catastrophes. Thus the concept of actionable transparency captures the extent to which everyone with an interest in improving the design has the right and means to act on both their own copy and the master copy of the design. pp. 28 - 29The concept is well captured in the conceptual form of Open Source code improvements in People, Ideas & Objects. The issue of scope and scale are addressed through the ownership of the intellectual property and the means to fund these developments. Otherwise I see no difference.
What does actionable transparency achieve that conceptual transparency cannot? As indicated at the beginning of the paper, design is a process in which people define problems and then draw on their stores of knowledge and generate new knowledge to solve those problems (Simon, 1981; Alexander, 1964). Much design-relevant knowledge is tacit and initially inaccurate (Nonaka, 1994), consisting of conjectures of the form: “if I change the design this way, these things will happen.” “This way” and “these things” are generally tacit hunches, which are not well-articulated even in the mind of the designer (Bucciarelli, 1994). Yet if a conceptually transparent design is also actionable, conjectures of this type can be tested, evaluated, and new conjectures generated quickly and efficiently. There is no need to make the conjecture comprehensible to another person. There is no need to persuade someone else that a new idea is worth trying, risking failure and embarrassment. Interactions between the designer and the design (embedded in a system of archives and test suites, etc.) are all that is needed to generate a new trial and new knowledge. p. 29In the Preliminary Research Report Professor Giovanni Dosi's work showed that innovation was messy. Many failures and wrong approaches would be taken, that it is this process of failure that helps to define the successful approaches.
Moreover, if a technical system is actionably transparent to several or many designers, experiments can go on in parallel and concurrently across many designers, who can then learn about and use each other’s changes via the system itself. (See Lakhani and von Hippel, 2009 on “optimistic concurrency” in open-source development.) This form of concurrent, recombinant experimentation creates a very rapid and powerful “generator-test” cycle. Such cycles, sometimes called “variation-selection -retention” cycles, lie at the core of all Darwinian evolutionary processes, including those found in theories of organizational change and evolutionary economics (Campbell, 1969; Simon, 1981, pp.128-130; Nelson and Winter, 1982; Anderson and Tushman, 1990; Tushman and Rosenkopf, 1992; Nelson, 1995). Thus actionable transparency can speed up the processes of design evolution, thereby increasing the rates of innovation and improvement for the system as a whole. p. 29It is necessary to state how the unique way in which the intellectual property of People, Ideas & Objects enables "Actionable Transparency". I as author of the original Preliminary Research Report, this blog and other writings earned the copyright to the ideas that are expressed. Users, the Community of Independent Service Providers and others are monetarily compensated for their contributions. In essence I am purchasing the ideas that are generated based on the original ideas expressed in the research and blog. The aggregate copyright of all of the ideas is then licensed to the users and members of the CISP's, developers and others, such that this type of interaction, experimentation and development is possible and I would suggest encouraged. I in turn use the copyright to generate the funding from the producers.
The cost of coordinating an organization or team is sometimes equated with its communication complexity, that is, the number of messages that must be passed between members in the course of getting the work done (Brooks, 1975). Mirroring and actionable transparency in a shared system have very different implications for communication complexity. Given a team of n agents working on a fully interdependent design, the cost of coordination via mirroring is n2 n. For the same number of agents and design structure, the cost of coordination via actionable transparency falls to 2n: each agent only needs to manage his or her information exchanges with the system. p. 31What is the effect of this conceptual model?
In the presence of actionable transparency, it is common for more complex organizational patterns to emerge in lieu of genuine mirroring. We describe these patterns below. p. 31If your an enlightened producer, an oil and gas investor or shareholder, who would be interested in funding these software developments and communities, please follow our Funding Policies & Procedures, and our Hardware Policies & Procedures. If your a government that collects royalties from oil and gas producers, and are concerned about the accuracy of your royalty income, please review our Royalty Policies & Procedures and email me. And if your a potential user of this software, and possibly as a member of the Community of Independent Service Providers, please join us here.
Technorati Tags: People's Baldwin Actionable-Transparency Innovation Research CISP User Wiki