This article was downloaded by: [132.211.1.

50] On: 31 August 2015, At: 20:14

Publisher: Institute for Operations Research and the Management Sciences (INFORMS)
INFORMS is located in Maryland, USA

Interfaces
Publication details, including instructions for authors and subscription information:
http://pubsonline.informs.org

ASP, The Art and Science of Practice: Elements of a

Theory of the Practice of Operations Research: Practice
as a Business
Frederic H. Murphy,

To cite this article:

Frederic H. Murphy, (2005) ASP, The Art and Science of Practice: Elements of a Theory of the Practice of Operations
Research: Practice as a Business. Interfaces 35(6):524-530. http://dx.doi.org/10.1287/inte.1050.0147

Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial use
or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher
approval, unless otherwise noted. For more information, contact [email protected].

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness
for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or
inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or
support of claims made of that product, publication, or service.

© 2005 INFORMS

Please scroll down for article—it is on subsequent pages

INFORMS is the largest professional society in the world for professionals in the fields of operations research, management
science, and analytics.
For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org
 informs ®

Vol. 35, No. 6, November-December 2005, pp. 524–530 doi 10.1287/inte.1050.0147

issn 0092-2102
eissn 1526-551X
05
3506
0524 © 2005 INFORMS

ASP, The Art and Science of Practice:

Elements of a Theory of the Practice of
Downloaded from informs.org by [132.211.1.50] on 31 August 2015, at 20:14 . For personal use only, all rights reserved.

Operations Research: Practice as a Business

Frederic H. Murphy
Fox School of Business and Management, Temple University, Philadelphia, Pennsylvania 19122, [email protected]

In the previous two articles in this series (Murphy 2005a, b), I looked at practice from the perspective of
doing operations research (OR). Because OR practitioners have careers and often found businesses, we need
to understand practice as a business. Operations research/management science (OR/MS) businesses face many
of the same issues as other technology and consulting businesses. However, the product is intangible, creating
marketing problems. The businesses operate in a complex ecology with different levels of competition and
cooperation among OR and software firms, depending on their different areas of expertise. Social networks play
an important role in the way individuals and companies market and develop new strands of practice.
Key words: OR/MS implementation; consulting; OR/MS practice.

O perations research (OR) practice is a business or

a part of business with the marketing and strate-
gic aspects of any industry. The literature specific to
He also gives three reasons why companies hire
consultants:
(1) They need outside expertise.
OR firms is small, but it provides anecdotal evidence (2) They have the expertise in house but are overex-
on how one can run an OR business (Corbett and tended and simply need more of it.
Van Wassenhove 1993, Corbett et al. 1995, Overmeer (3) They have the expertise in house, but the per-
et al. 1998, Ormerod 2002). Ormerod (2002) provides son hiring the consultant doesn’t want to use it for
an extensive bibliography. political reasons.
The business is divided into three broad areas, sell- Central to success in consulting is to move rapidly
ing tools, developing decision-support systems, and along the learning curve to develop institutional or
consulting (external and internal). Firms can operate software expertise that lowers the cost of meeting the
in more than one area. However, they have to be customer’s needs. Developing expertise, along with
mindful of their relationships with other firms. developing a reputation among clients, becomes a
Consulting is often the starting point for an OR strand of practice (Corbett et al. 1995).
business, and companies that sell decision-support
systems also provide consulting services to their
clients. Consultants typically specialize by industry or Relationships Among OR Businesses
tool if they are boutique firms or run a practice within Firms in the OR business operate in a complex eco-
a larger firm. Theurer (2000) describes three levels of logical environment. They have to decide who their
consulting: competitors are and who their complementors are
(1) Management consulting, an advisory role re- (firms that have complementary skills or products that
quiring a high level of management-consulting skills; are part of delivering customer services). Firms sell-
(2) Large-project consulting, a team leadership role ing decision-support tools in the same market are
requiring project-management skills; and competitors. The companies that provide embedded
(3) Individual problem solving, a technically ori- solvers are complementors to firms that sell matrix-
ented role requiring strong quantitative problem- generation software. In the information-technology
solving skills. industry, IBM sells middleware and consulting
524
 Murphy: ASP, The Art and Science of Practice
Interfaces 35(6), pp. 524–530, © 2005 INFORMS 525

services. Developers of end-use applications use IBM The decision-support system (DSS) developer gets to
middleware in their products. choose the solver, not the developer of the model-
A variety of components can go into a decision- ing language. A supplier of modeling languages has
support system; the components can come from sev- to accommodate all of the major solvers because dif-
Downloaded from informs.org by [132.211.1.50] on 31 August 2015, at 20:14 . For personal use only, all rights reserved.

eral different companies. In an optimization-based ferent solvers have different strengths and customers
decision-support system, OR firms provide the mod- often have to tailor the solver to the problem. Whether
eling language and solver and the complete package a solver company should enter the market depends
that the customer sees as the decision-support sys- on profitability. Sellers of modeling languages are
tem. Software firms provide the database and inter- generally very small, indicating that the market is
face tools, which are generic to many applications small with low barriers to entry, which implies that
(Figure 1). profitability is also low. Concomitantly, developers of
The relationships among the firms can become com- modeling languages and solvers have to be careful
plicated and depend on the markets of each of the about engaging in consulting. Consultants are their
firms. The OR market is a small source of business best customer, and they cannot compete with them
in user interfaces. So, the firms that provide inter- without the risk of losing their tools business.
face software do not pay much attention to this mar-
ket, and they are not a competitive threat to an OR
business. However, they can make decisions that turn
Life Cycles of Applications
an OR firm into collateral damage. For example, the Applications areas have natural life cycles like any
major problem with using spreadsheet optimization products. When a model has a significant impact in
tools is that spreadsheets do not represent sets or set a new area, the model becomes a strategic asset, and
operations. When Lotus Symphony added a represen- the inventing organization keeps knowledge about it
tation of sets to spreadsheets, one company decided as an internal trade secret to the extent possible. The
to use Symphony as the basis of a matrix generator idea eventually leaks out, and others develop similar
and model-management tools. When Lotus decided models. After a while, the core ideas of the appli-
to discontinue the product, the company was driven cation become common knowledge and secrecy is of
out of business. no value. The usability of the model as embedded
A firm that moves into the business of its customers in a decision-support tool becomes as important as
faces the risk that its other customers will bolt, think- the model itself, and the market value lies more in
ing they will be next. For example, the electronics the software than in the idea. The costs of doing all
assemblers Solectron and Flextronics will never pro- of the decision-support development are best spread
duce their own branded products because all of their over several customers. At this time, the company
major customers would then abandon them. that developed the model internally can spin it off or
The level of risk a firm takes depends on the free- can replace the internally developed model with one
dom of action of its customers. A solver company developed and maintained externally. This is what
can enter the modeling-language business without firms have done with refinery and yield-management
much risk because of the structure of the market. models. The life cycle of an application shapes the
industrial organization of the business.
The life cycle has to do with the economics of
Database Modeling Decision- knowledge. An idea can be costly to develop and
tools language support Customers
system require much trial and error. However, it is cheap
Solver User to duplicate. Recognizing this progression, nations
interfaces
tools developed patent law to protect intellectual property.
Recently, patent law has been extended to software
and to models and algorithms embedded in software.
Figure 1: A knowledge chain of software tools are included in deliver-
ing an optimization-based decision-support system. The components can To some extent, the ability to patent models
come from a variety of companies. (Murphy 2002) can alter their life cycles. For example,
 Murphy: ASP, The Art and Science of Practice
526 Interfaces 35(6), pp. 524–530, © 2005 INFORMS

under current US law, American Airlines might have Most ongoing consulting firms rely on extend-
been able to patent the idea of using software tools ing their existing practice areas, yet businesses have
for yield management as a general concept. American to start somewhere. Typically, these businesses are
Airlines would then have had proprietary use for a started by people with experience and connections to
Downloaded from informs.org by [132.211.1.50] on 31 August 2015, at 20:14 . For personal use only, all rights reserved.

fixed number of years, giving it a competitive advan- their initial clients. Beam (2004) describes how she
tage for many years, creating opportunities for the is developing a practice as a consultant and indi-
company’s consulting arm in a wide range of busi- cates the books she has found useful. She covers
nesses and providing the company with licensing rev- how she does cold calling. Her experiences reflect a
enue. Such patent laws also would have slowed the true startup with no initial client list. Theurer (2000)
models’ growth into other application areas. planned to leave a corporate job once he had lined
How patent law will change the business is up his first client. In the end, he quit first after six
unknown. First, the enforceability of broad patents months of looking for the first client without suc-
in business processes is unclear. Second, if an author cess. Unlike Beam, he has seen no correlation between
presents research in a public forum and does not marketing efforts and projects. Yet, he has used his
apply for a patent within a specified time, the knowl- marketing training to close deals and develop and
edge is freely available. Thus, fundamental ideas maintain clients. Both Beam (2004) and Flowers (2004)
developed in a university that are not part of an offer advice on how to close contracts and get the
immediate application will likely not get patented work done.
because of the length of time from idea to product.
I have to qualify this statement because universi-
ties can now patent the products of government- Business Networks
sponsored research and are looking to patents as a Ormerod (2002) points out that the main clients for
new source of revenue. OR models are large organizations and that OR func-
tions as a change agent in these organizations. He also
Entering the Market notes that the structure of the OR business falls in
Companies, broadly speaking, are not willing to pay the subject area of industrial organization and cites
for general OR skills if they know of a DSS they can Abbott (1988) as a source for understanding the soci-
acquire or a specialty firm practicing in the area of ology of professions, which affects the structure of the
the business problem. In the case Corbett et al. (1995) market. Because the field of OR has no method for
discuss, the company they studied was in the right accrediting practitioners, Abbott calls it a semiprofes-
place when pension laws in the Netherlands changed. sion, which means that no group controls entry, unlike
When they wrote their article, the firm was also try- the certifying boards for doctors and state bar associ-
ing to convert models from projects into decision- ations for lawyers.
support systems, as its project business was getting Sociology is relevant in another way. As OR work is
squeezed out of consulting areas in which competi- typically sold through referrals, an essential feature of
tors had developed decision-support tools. understanding the practice business is the structure of
Seemingly, the main ways to enter a market are the social networks that enable a business to succeed.
(1) to work for a client with a need in a technol- Three social networks are involved in practice. One
ogy or a subject area with no existing players, (2) to allows practitioners to link up to form what Ormerod
serve a market in which a captive consultant serves calls a web of practice, which brings different kinds
only one player (the broadening of the original yield- of OR expertise together to do a project. The second
management work), (3) to use sweat equity to get is the social network that connects practitioners with
an initial version of the software (Maximal Systems), clients. The third is a social network linking academia
(4) to obtain government research funding of the first with practice.
version of the software (ILOG interior-point solver), Corbett et al. (1995) describe a firm in which the
and (5) to extend existing products or services to new social network of the firm consisted of graduates
areas. from the same university. Other networks develop as
 Murphy: ASP, The Art and Science of Practice
Interfaces 35(6), pp. 524–530, © 2005 INFORMS 527

clients refer colleagues to the practitioners, a method often an initial customer for important new applica-
better known as word of mouth. As long as the OR tion areas.
consulting firm maintains an understanding of its Large organizations with immediate needs also
client’s firm and maintains relationships with the peo- fund development projects. The modeling language
Downloaded from informs.org by [132.211.1.50] on 31 August 2015, at 20:14 . For personal use only, all rights reserved.

ple in that firm, the client tends to provide a stream GAMS was developed to meet the needs of the World
of business. Understanding these social networks and Bank, and then the developers founded a company.
their value in marketing is crucial to understanding Large private organizations or consortia can also fund
how to develop an OR business. development. AMPL was developed at AT&T, which
Social networks are especially important for solo then licensed it out once it saw that the market for
practitioners. Beam (2004) describes the ones she has the product was not big enough to interest a large
developed. Theurer (2000) uses networks of business company. After leaving Exxon, Thomas Baker put
contacts and receives calls from people in his and together a consortium of companies to fund the devel-
his clients’ social networks. He keeps his network opment of the MIMI optimization-modeling system.
fresh through regular interactions with his collection Consortium members also formed the main customer
of clients. He has an interesting rule: he never takes base. He regularly turned to these companies to fund
a long-term job because he would become too depen- further development of the system. (He eventually
dent on that one client, his networks would age, and sold his company to Aspen Technologies; OR busi-
he would stop getting calls. Essentially, he uses a port- nesses are subject to the same merger-and-acquisition
folio strategy to minimize the risk of running out of dynamics as other companies.)
customers. In the entrepreneurship literature, one category of
The network that is important in cutting-edge areas start-ups is the “garage” startup. The founders of such
companies start their businesses by working nights
is that between academics and practitioners, with
and weekends without any formal business struc-
some academics also being practitioners. Professional
ture or external investment. Bjarni Kristjansson devel-
societies, such as INFORMS, provide this social net-
oped Maximal Software’s MPL modeling language
work. The relationship between academics and practi-
this way.
tioners varies by profession. Academic research tends
Although researchers have done some research on
to drive new OR technologies. In information sys-
the structure of the practice business and the whole
tems, companies set most of the research agendas and
ecology of institutions that create new ideas and keep
academics follow what is happening in practice, the
useful ones circulating, much more work is needed
reverse of OR. However, OR practitioners drive mod-
to fully understand the industrial organization and
eling issues and new areas of application. This means
sociology of the field.
that academics and practitioners are coequals in gen-
erating new ideas and research directions (Schon
1983).
Conclusions for the Three Papers in
the Series
The different approaches to understanding problems
Starting a Tools Business presented in the literature are really specializations
To understand the industrial organization of the tools of a basic heuristic for learning and doing. How-
business, we must include the government. Organiza- ever, as Ernst and Newell (1969) learned, a general
tions generally need initial funding to cover the devel- problem solver is insufficient for all aspects of prob-
opment costs of tools. Yet, I have seen little venture lem solving. The semantic nets for different subject
capital funding tools firms. The government has been areas are very different because the subject areas have
the primary source of funding for optimizers. To get unique features. Woolsey’s (1998) focus on processes
the word out, the US government funded Charnes and his emphasis on making sure what the process
and Cooper’s initial work in developing linear pro- is by doing it, is very different from the soft OR
gramming applications in companies. The military, practitioners’ seeking the perceptions of many differ-
because of its mission, scale, and use of technology, is ent people to understand the broader organizational
 Murphy: ASP, The Art and Science of Practice
528 Interfaces 35(6), pp. 524–530, © 2005 INFORMS

issues. A full-blown soft OR analysis, however, is too and conducting computer simulation studies of pro-
resource intensive for most problems even though it cesses constitutes the largest class of opportunities
is a useful framework for understanding the issues for OR with the largest total benefit. The opportu-
of implementing models in organizations. Further- nities should be greatest in working on processes
Downloaded from informs.org by [132.211.1.50] on 31 August 2015, at 20:14 . For personal use only, all rights reserved.

more, we can understand the relationship between because businesses have more processes than systems
the modeling- and algorithm-development branches and new technologies create new opportunities to
in OR/MS by noting the commonality in the heuris- redesign processes. Nevertheless, the market for opti-
tics used to solve problems but the very big differ- mizing systems is real and large, and the problem of
ences in the subject-area semantic nets. understanding complex organizations and social sys-
Because the 18 problem types (Murphy 2005a, tems is also real.
Table 1) cover too broad a range for any one person I have documented some of the areas of technical
to be expert in all areas, we should learn enough from expertise and largely ignored the management, com-
others to be able to work effectively on problems that munication, and marketing skills. Project manage-
fall on the boundaries separating our expertise from ment is widely discussed by experts in that subject.
theirs rather than ignore their special expertise. We Communications and marketing are research topics
need a team approach and team skills to complete OR in other fields. Our profession is unusual in that our
projects successfully. products are intangible. Just as Tufte (1983) has devel-
Each subject area has its own semantic nets and oped ways to communicate data visually, we need to
procedures. Builders of linear programming solvers develop methods to make mathematical models tan-
gible. Visual simulation tools have a strong impact
carry a large semantic net of numerical methods.
on managers because the computer-generated move-
Business analysts who focus on corporate decisions
ments make the mathematics concrete. Spreadsheets
have semantic nets based on their experiences making
are ubiquitous because they are algebra by exam-
corporate-level decisions or observing the decisions of
ple and communicate to managers who have never
others and their consequences. OR practitioners in the
understood the meaning of x in algebra. Doing the
airline industry carry a semantic net on the structure
equivalent for representing mathematical programs
of that industry (Corbett et al. 1995).
could have the same impact on their use.
What I have labeled as a universal heuristic is really
We need to understand what personality traits,
a checklist for developing a subject-area heuristic.
skills, and organizational positions make people effec-
This checklist is the basis of a metaheuristic for for- tive champions in an organization on the user side of
mulating the subject-specific heuristics. modeling, especially for breakthrough models. What
The issues I have discussed raise many issues that made Robert Crandall such an effective user of mod-
could form a productive research agenda, especially eling at American Airlines? Other CEOs followed his
for OR faculty in business schools who can team up lead only after they saw that modeling had clear
with colleagues in the social sciences. Can we doc- benefits. Probably the best champions are “intuitors”
ument expertise in OR/MS? Can we synthesize this who can intuit possibilities on the Meyers-Briggs
expertise to help people acquire expertise and learn measures of personality. Most people are “sensors”
faster? Can we evaluate the relative merits of dif- and the opposite of intuitors. I suspect that sensors
ferent approaches? Can we document the nuances are more amenable to using existing models than to
of practice? How can we organize the intellectual developing new ideas. We need techniques that help
and business aspects of OR/MS practice? What busi- us to communicate to the sensors. Developing these
ness strategies succeed? What chain links technical techniques would make for a research career com-
innovation to practical implementation? What are the parable to Tufte’s career in analyzing and improving
social networks and ecology of the field? How can we data presentation.
improve the networks? One of the big problems facing practitioners is
Where can we make the biggest improvements achieving the first step, getting some manager to rec-
in organizations using OR? My suspicion is that ognize that OR can solve the problem. The manager
focusing on basic processes, as Woolsey recommends, has to make a successful analogy between his or her
 Murphy: ASP, The Art and Science of Practice
Interfaces 35(6), pp. 524–530, © 2005 INFORMS 529

problem and some problem with an OR solution. As the ethics of the way in which the mayor’s office in
discussed earlier, analogical processes are central to New York City used analysis results to shape poli-
the successful design and construction of OR mod- cies affecting the fire department. How should ethical
els. The difficulty is that superficial similarities are issues be addressed?
Downloaded from informs.org by [132.211.1.50] on 31 August 2015, at 20:14 . For personal use only, all rights reserved.

easier to see than the deeper, structural similarities Standard research methods are inappropriate. Case
of OR models. Furthermore, managers, like everyone, studies are good stories but may be bad science. The
tend to draw upon problem-solving measures they social-science model with its rigorous data-collection
have used successfully in the past. This tendency con- and statistical methods straightjackets thinking and
tributes to the pattern of OR successes clustering in can lead people to work on problems that are side
particular industries: managers in those industries are issues rather than central issues in a field. Argyris and
used to using OR, and superficial similarities often Schon (1996) offer a template for action research as
match deep structures. If we could develop methods a useful alternative. In action research, the researcher
for helping managers to recognize the deep structures either works with a practitioner or becomes one and
of their problems, they might think of OR solutions then reflects on his or her experiences and analyzes
more often. An interesting attempt to help managers them. Essentially, analysts developed soft OR in that
recognize whether their problems are amenable to OR way. No one rigorously tested the methodologies. Yet,
techniques is INFORMS’ Science-of-Better campaign users of the approach believe that they work.
(www.scienceofbetter.org). The Web site offers them We do not need to consider average practice be-
advice on when OR techniques can make a difference. cause we want to improve the best practices. We are
A valuable addition to the field would be a formal concerned with best practices and with accumulating
analysis of the industries that have used OR success- knowledge that leads to expertise, as was Armstrong
fully and their characteristics. (2001) who sought to elucidate the principles of fore-
One of the critical issues for OR practitioners, casting. One way to capture this knowledge is to use
but not OR researchers, is learning how to work the interactive capabilities of the Web to build a set
with clients and colleagues. Working with clients of principles, starting with a seed based on an infor-
successfully requires what is now termed emo- mal set of communications and an initial synthesis of
tional intelligence. Kanazawa (2004) argues that gen- the literature on the craft of OR, taking the title from
eral intelligence, and OR professionals must be highly Miser’s (1992) paper.
intelligent to understand the subject, is actually a Carefully examining the nonmathematical side of
narrow adaptation that allows humans to adapt our field can enrich OR practice and should lead to an
to changes in circumstances. He uses research by “epistemology of OR/MS practice” (Miser 1998, p. 36).
Derksen et al. (2002) and Davies et al. (1998), who
found essentially no correlation between general intel-
ligence, which we think of as IQ, and emotional intel- References
ligence, the ability to appraise the emotions of others Abbott, A. 1988. The System of Professions: An Essay on the Division
of Expert Labor. University of Chicago Press, Chicago, IL.
and express our own emotions. (I could use these Argyris, C., D. A. Schon. 1996. Organizational Learning II. Addison
findings to explain the strange behaviors exhibited by Wesley, Reading, MA.
faculty colleagues.) I do not know if a literature exists Armstrong, J. S., ed. 2001. Principles of Forecasting: A Handbook
for Researchers and Practitioners. Kluwer Academic Publishers,
on emotional intelligence and success in consulting. If
Norwell, MA.
not, we need one. Beam, C. 2004. ASP, the art and science of practice: How I started
A project has to succeed both technically and orga- an OR/MS consulting practice with a laptop, a phone, and a
nizationally. What technical and organizational ele- PhD. Interfaces 34(4) 265–271.
Corbett, C. J., L. N. Van Wassenhove. 1993. The natural drift: What
ments are critical to success? Can we assess these a happened to operations research? Oper. Res. 41(4) 625–640.
priori to increase our chances of success and reduce Corbett, C. J., W. J. A. M. Overmeer, L. N. Van Wassenhove. 1995.
the odds of failure? What methods can we use to man- Strands of practice in OR (the practitioner’s dilemma). Eur. J.
Oper. Res. 87(3) 484–499.
age technical risk in projects? One last question: What
Davies, M., L. Stankov, R. D. Roberts. 1998. Emotional intelligence:
are the ethical dimensions of OR practice? Green and In search of an elusive construct. J. Personality and Soc. Psych.
Kolesar (2004) describe an experience they had with 75(4) 989–1015.
 Murphy: ASP, The Art and Science of Practice
530 Interfaces 35(6), pp. 524–530, © 2005 INFORMS

Derksen, J., I. Kramer, M. Katzko. 2002. Does a self report mea- Murphy, F. H. 2005a. ASP, the art and science of practice: Elements
sure for emotional intelligence assess something different than of a theory of the practice of operations research: A framework.
general intelligence? Personality and Individual Differences 32(1) Interfaces 35(2) 154–163.
37–48. Murphy, F. H. 2005b. ASP, the art and science of practice: Elements
Ernst, G. W., A. Newell. 1969. GPS: A Case Study in Generality and of a theory of the practice of operations research: Expertise in
Downloaded from informs.org by [132.211.1.50] on 31 August 2015, at 20:14 . For personal use only, all rights reserved.

Problem Solving. Academic Press, New York. practice. Interfaces 35(4) 313–322.
Flowers, A. D. 2004. ASP, the art and science of practice: Getting Ormerod, R. J. 2002. On the nature of OR: Taking stock. J. Oper. Res.
the R in contact. Interfaces 34(5) 377–382. Soc. 53(5) 475–491.
Green, L. V., P. J. Kolesar. 2004. Improving emergency respon- Overmeer, W. J. A. M., C. J. Corbett, L. N. Van Wassenhove.
siveness with management science. Management Sci. 50(8) 1998. Cooperation between strands of practice: Challenges and
1001–1015. opportunities for the renewal of the profession. J. Oper. Res.
Kanazawa, S. 2004. General intelligence as a domain-specific adap- Soc. 49(4) 369–380.
tation. Psychological Rev. 111(2) 512–523. Schon, D. A. 1983. The Reflective Practitioner: How Professionals Think
Miser, H. J. 1992. Craft in operations research. Oper. Res. 40(4) in Action. Basic Books, New York.
633–639. Theurer, G. J. 2000. Lone Ranger rides again. OR/MS Today 27(1)
Miser, H. J. 1998. The easy chair: Judging effectiveness of work 28–31.
in OR/MS practice and claiming credit for its consequences. Tufte, E. R. 1983. The Visual Display of Quantitative Information.
Interfaces 28(4) 31–37. Graphics Press, Cheshire, CT.
Murphy, F. H. 2002. The occasional observer: www.uspto.gov, A Woolsey, R. E. D. 1998. The fifth column: On doing well by
new source for what is happening in operations research prac- doing good and an offer of free education. Interfaces 28(2)
tice. Interfaces 32(3) 26–29. 99–103.