Eric V. Price
Lexis-Nexis
9595 Springboro Pike
Dayton, Ohio 45342
Tel: (937) 865-6800
Fax: (937) 865-1655
Email:
Eric.Price@Lexis-Nexis.com
This position paper presents a taxonomy of behavioral and cultural factors which influence an organization's ability to adopt effective reuse practices. This demonstrates that impediments to effective reuse are deeply rooted in human nature, and can best be overcome through an insightful, holistic, multidisciplinary effort.
Through awareness of sociological components of an organization's culture and the psychological components of the behavior of the individuals within the organization, reuse advocates will have clearer understanding of impediments to a reuse program, direction for corrective action, and a rationale for allying with other proponents of organizational development (such as the corporate human resources department). Synergy with corporate mandates such as diversity and multiculturalism-when done well-can contribute to the organization's psychological well-being (reality testing and inclusiveness) and increase the likelihood of adopting effective reuse practices.
Keywords: Organizational culture, corporate ethnography, sociology and psychology of software development.
Workshop Goals: Promoting an interdisciplinary understanding of behavioral impediments to software reuse; reality testing the positions taken in this paper, finding kindred spirits, identifying ``secure bases'' that facilitate software reuse (such as a reuse handbook).
Working Groups: Domain Engineering Helps Manage Change - Hype, Myth, Wishful Thinking, or Reality? Object Technology, Architectures, and Domain Analysis: What's the Connection? Is There a Connection?
In deploying technically successful reuse programs at four sites, I have observed a number of differences which contributed to the sustainability of the reuse program and to the future of the site itself. Most recently, a corporate reuse program contributed to the site's successful deployment of a ``multimillion dollar overhaul of its infrastructure'' which is itself a reusable framework allowing the company to build new products which ``maintain its status as one of the premier information services.''
These sites differed significantly in a number of organizational culture dimensions that either directly or indirectly affected the continuity of the reuse effort. Many of these are equally likely to be the subject of an entire book, or be dismissed as a mere management fad: the degree of management awareness of the company's own corporate culture; the project models (monolithic versus silicon valley); the degree of customer focus and contact; the relevance of internal measurement systems to customer-derived product requirements; the willingness and ability to change business practices based on industry benchmarking or external audits; the degree of management awareness of reuse efforts, the degree of management acknowledgement of the success of reuse efforts; the cost-effectiveness of the reuse efforts; the presence and use of a formal requirements elicitation processes; the training of software developers (software development tools and methodologies, project management, process management, teamwork, financial awareness); the degree of autonomy and empowerment enjoyed by the development staff; and the breadth of responsibility as well as the exposure of engineers to the strategic long-range plan.
In all cases, software reuse was a technical success. Assessing the overall success of the reuse efforts is complicated, however, by closure of the sponsoring program or plant, attrition of the reuse staff, and dismal failure of some of the applications for which the reusable components were originally intended. Moreover, significant behavioral issues impacted deployment of reusable components. In fact, the customers themselves were deeply conflicted over issues such as: the compatibility of a library's internal consistency checks with the letter of an industry standard; error handling (exceptions versus return codes); and use of the C preprocessor (``bizarre'' versus ``best alternative''). One customer chose to ignore the existence of a new library due to the volume of new application development (i.e. workload) it made possible.
Ultimately, the reuse programs (and the organizations which sponsored them) succeeded or failed as a result of the aggregation of individual behaviors. While these behaviors were impacted through initiatives such as performing a Malcolm Baldrige Award Assessment, the behaviors were rarely explicitly acknowledged or systematically addressed. Individual behaviors which can overwhelm a reuse program can, in larger concentrations, gain the attention of social scientists. Thus, the following social sciences can give insight into successful software reuse.
As primates, human males are predisposed to form dominance hierarchies; all humans are predisposed toward mimicry [1]. The result can be unhealthy competitiveness, authoritarianism, abuse of symbolic status [2], and groupthink [3].
The human predispositions toward hierarchy and mimicry create needs for compliance and conformity. These were studied prior to World War II by German sociologist Max Weber, a proponent of bureaucracies and charismatic leadership. More recent work has identified four motivators common within organizations: affiliation (loyalty cliques), personal (positional) power, institution-building power, and achievement [4]. Because the ``institutional manager'' prefers activity to reflection, but the ``achievement manager'' abhors centralization, a hybrid of these two would be most likely to support enterprise-wide reuse of a carefully designed and thought-out software framework.
At the end of World War II, Christian and Jewish organizations in the U.S. chartered behavioral scientists to develop ``a clear understanding of how people of different races, colors, and creeds can live and work together, and how best we can achieve harmony'' [5, 6]. The insights into the inter-personal and communication dynamics from this work also apply to the diversity of cognitive and personality types necessary for good teamwork among intellectual property creators.
Optimizing organizational teamwork is addressed in the seminal work on team typology by Meredith Belbin which provides descriptions of roles (e.g. ``director,'' ``resource investigator,'' ``monitor-evaluator,'' etc.) and organizational contexts in which they best interact.
Peter Senge argues that organization learning is facilitated through the five disciplines: self mastery, mental modeling, building shared vision, team learning and systems thinking [7]. These disciplines are core competencies of engineers but are more likely to be infused in engineering school than consciously taught. In particular, it is self mastery that allows maintenance of the ``creative tension'' between the vision and current reality.
In [8] Stephen Covey elaborates on Senge's ``self-mastery:'' be proactive, begin with the end in mind, put first things first, think win-win, seek first to understand, synergize, and sharpen the saw. Covey states that trust can be built only through trustworthiness, which requires development of both competence and character. Trust is critical for overcoming the ``Not Invented Here'' syndrome which often impedes reuse.
Much of the improvement in American manufacturing competitiveness is due to improved understanding of motivation, such as [9]. Reuse programs based on extra incentive pay for producing reusable code provide a hygiene factor (e.g. more stuff) when an intrinsic motivator (e.g. Deming's pride of workmanship) may be more appropriate.
The European management guru Manfred Kets de Vries presents several taxonomies of business dysfunction, such as paranoid, depressive, and compulsive organizational cultures [10]. His work provides invaluable insights for the reuse practitioner, including the observation that empathy is a prerequisite to trust [11].
A more accessable application of clinical psychology to the workplace is found in [12], which maintains that key indicators of psychological well-being are reality testing and inclusiveness.
The Kohlberg Moral Maturity Model provides developmental insight into how one's notion of justice impacts relationships, as well as an individual's ability to be truly empathetic. Moreover, the highest Kohlberg levels (``post-conventional morality'') enable the conflict resolution and interface contract negotiation required for creating ``well-behaved'' software components and frameworks.
In his excellent discussion of learning in the software industry, Marc Sacks notes the otherwise limited applicability of educational psychology due to its assumption that learning occurs primarily in the classroom [13]. The case study presented by Sacks illustrates many of the practical difficulties successful software reuse must overcome.
To the extent that software reuse promotes ``the common good,'' it impinges on the field of moral theology. The workplace applications of moral theology noted by [7], [14], [15] and [8] are treated in detail by [16]. The literature of moral theology deepens the appreciation for the profound conversion which the reuse evangelist seeks to achieve.
Attributes such as trust, empathy, pursuit of common good, inclusiveness, and reality checking are important for the proper functioning of any entity made up of people. This is especially true for a software reuse group which hopes to serve a large number of developers who in turn must balance demanding requirements. Taken together, these attributes contribute to the psychological well-being of an organization. Weighed individually, however, it is reality testing and inclusiveness that are most easily observed and objectively assessed.
The following propositions are organized into two principle themes, first the importance of general behavioral issues in encouraging software reuse, and second, issues relating to the psychological well-being of the organization which will impact software reuse.
Proposition 1: Reuse is primarily a behavioral issue.
The business rationale for software reuse (at least as a ``component industry'' [17]) has been documented for thirty years. ``Don't reinvent the wheel'' (software reuse) is as easy to grasp as ``don't waste time on non-value-add activites'' (reengineering) and ``do it right the first time'' (Total Quality Management). Slowness in adopting such an obvious windfall-an apparent denial of common sense-is explained in Scott Adams' The Dilbert Principle: ``It's useless to expect rational behavior from the people you work with, or anybody else for that matter.''
Software technologists are reluctant to conclude that such an important part of the future of software technology is influenced primarily by non-technical issues; In 1990, Bertrand Meyer announced that new technologies heralded ``a shift to a `new culture' whose emphasis is not on projects but instead on components'' [18]. Unlike Meyer, the world's leading high-tech CEO's are not so sanguine in assuming improved technology will induce cultural change: ``changing cultural values is really a company heart transplant-a delicate operation in which people who know what they are doing proceed with surgical precision, clear understanding, and with profound respect for the risks and rewards involved [19].'' Similarly, CEOs from HP report spending ``a lot of my time talking about values rather than trying to figure out the business strategies'' and feeling like a ``spiritual'' leader.
Proposition 1.1: The cultural shift toward software reuse is a multidisciplinary organizational transformation.
In [20], the leading creator/practitioner of Total Quality Management discusses the components of TQM that must all be addressed for success: leadership, commitment, product, process and organizational structure. Software reuse programs must inter-weave these same five components. While process is most often addressed, product has recently received increased attention. Organizational structure has been studied for high-performance software development groups (although not specifically in the context of software reuse) [21]. Finally, behavioral issues (leadership and commitment) in software reuse are often acknowledged, but direct application of the behavioral sciences to software reuse is rare.
A similar example is found in the context of re-engineering, where one of the founders of reengineering, Michael Hammer, recently admitted, ``I was reflecting my engineering background and was insufficiently appreciative of the human dimension. I've learned that's critical.''
Proposition 1.2: Intrinsic motivators are most effective in promoting software reuse
Extrinsic motivators are unlikely to compensate for the considerable effort involved in creating reusable software. Thus, intrinsic motivators such as peer recognition and pride of accomplishment should be fully acknowledged and enabled as important motivations for creating reusable software. This is not only more cost-effective and likely to succeed than extrinsic motivators such as monetary rewards, it is also less likely to become a target of cost-cutting. Note that the intrinsic motivators must be as real and substantial as extrinsic motivators: the organization must have the credibility and steadfastness to provide genuine psychic rewards.
Proposition 1.3: The cultural shift toward software reuse requires a ``secure base.''
A secure base lessens the apprehension associated with change. It could take the form of a leader who encourages innovation and experimentation, or a support organization that will promptly help overcome difficulties associated with new techniques. Most often, a commercial engineering firm's commitment to a product line provides the secure base for organizational change efforts. Because a software reuse program may require changes in commitment, process, and organization, the TQM components of leadership and product are better candidates for the secure base.
The subtlety of relationships in the change effort is expressed by [22]: ``The work of change belongs to organization members. The change agent's work is to give those members a place to feel secure while they go about their growth and development, without taking that discovery and learning process from them.''
Proposition 2: Exemplary psychological well-being within the organization is a key enabler of software reuse.
Clearly, the organizational neuroses described by [10] would pose an obstacle to software development of any sort. But the popular notions of ``programming by contract'' and ``egoless programming'' so important for software development in general and essential for the development of reusable software require a degree of psychological well-being that is nearly contagious. Most organizations would dismiss such a goal as wishful thinking, but some of the largest software development organizations at least acknowledge the ``rich psychocultural extravaganza, in which creativity, group dynamics, crude instincts, and technical fashion run riot'' [23].
Proposition 2.1: The cultural shift toward software reuse correlates with inclusive behavior.
Inclusiveness is demonstrated by participative management, employee empowerment and initiative and diverse representation on teams. These correlate to higher Kohlberg levels, and, in turn, with greater degrees of empathy and justice. Indeed, [24] maintains that the attaining post-conventional Kohlberg levels is essential for organizational competitiveness.
The difficult design, implementation, and support issues presented by software reuse can rival issues of racioethnicity, gender, and religion as litmus tests for discussability, mutual respect, and inclusiveness. Organizations which deal successfully with the latter issues are better positioned to deal with reuse issues.
Proposition 2.2: The cultural shift toward software reuse is facilitated by systematic corporate reality testing.
Reality testing is manifested by independent audits (such as SEI and Malcolm Baldrige National Quality Award), employee effectiveness surveys (that are taken seriously, i.e. acted upon), various activities to infuse the ``voice of the customer,'' and various professional collaborations. Inclusive behavior is often a prerequisite for reality testing to succeed.
Examples include Raytheon, which has achieved high levels of software reuse along with SEI recognition. Participation of the software engineering staff in these efforts was recently identified as the principal success factor [25].
Another example is AT&T, the winner of three Malcolm Baldrige Awards. AT&T has provided leadership in creating languages which enable reuse (e.g. the interoperability of C and C++) as well as reusable frameworks.
Inclusiveness and reality testing are not merely important skills used by developers of reuseable software, they are important environmental components as well. These are fostered by the selflessness and ``keen sense of justice'' manifested by ``institutional managers'' and the secure base of true meritocracy provided by ``achievement-oriented managers'' [4]. Such an organizational culture is able to motivate developers with true psychic rewards.
Systematic discussion of the implications of organizational culture behavior on software reuse are rare. One attempt merely adopts vocabulary from anthropology while presenting several reuse methodologies. More work would be valuable, such as investigating the difference in motivation between framework developers and application developers, extending the research in motivational differences in subgroups of information systems professionals in [26].
Eric V. Price is a senior software engineer at Lexis-Nexis where he provides support and mentoring for reusable C++ software, including HP's OODCE client-server framework. He holds a MS in mathematics from the University of Illinois and has previously worked in technology transfer as the liaison from NCR to the Microelectronics and Computer Technology Corporation.