Eric Peck, Lynndee Kemmet, Ray McGowan,
C. Reed Hodgin, and Bart Peintner
The Agility Imperative
Emerging Knowledge Management Requirements
for Stability Operations in the U.S. Army
David Kilcullen said it well in The Accidental Guerilla, when he argued that we
have entered a new era of warfare in which the “enemy” is often not nation-states
but non-state actors who are far more agile and adaptive than large nation-states.1
因此, succeeding in this era of warfare will require us to become more agile
as well so that we can adapt quickly to changing environments.
我们. soldiers and their commanders are aware of this need to be flexible. 在
both Iraq and Afghanistan, they have witnessed firsthand the enemy’s ability to
adapt quickly to changing circumstances in order to overcome efforts to defeat
他们. Insurgent groups have shown themselves to be mobile, capable of using sim-
ple technology to thwart sophisticated Western technology, and adept at quickly
using what information they gather to guide their actions. Unburdened by any
rule-bound system of how they must acquire and use resources, insurgent groups
are often one step ahead of coalition forces.
As in any conflict, one of the most valuable assets is information. 军队
forces of most nations, including the U.S., are well aware of this, and the collection
and management of information has long been considered critical for the success-
ful planning and execution of missions. Advances in technology have made it pos-
sible for modern militaries to collect a mass of data, but process has not kept pace
with technological advancement.
Despite the sophisticated new technology available to coalition forces in Iraq
and Afghanistan, insurgents still often seem to know things sooner and to act on
Brigadier General Eric Peck is Commander of the Kansas Army National Guard.
Lynndee Kemmet is a researcher and project manager for the ADT project at the
Network Science Center at West Point.
Ray McGowan is a Senior Engineer at the U.S. Army Communications-Electronics
Research Development Engineering Center.
C. Reed Hodgin is President of AlphaTRAC, Inc.
Bart Peintner is a Senior Computer Scientist at SRI International.
© 2012 Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart
Peintner
创新 / 体积 7, 数字 1
91
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart Peintner
that information faster than the coalition forces fighting them. Collecting informa-
tion has become much easier, but the “actionable” use of that information remains
a challenge. Soldiers and commanders are now overloaded with data, and sifting
through it to find what is and is not of value at any given moment is difficult.
而且, sharing information remains a challenge, partly because of technologi-
cal barriers between systems and partly because of barriers created by informa-
tion-sharing policies.
Civilian and military researchers are aware of these challenges and of the
urgent need to overcome them. Millions of dollars are poured into efforts to find
technical solutions to information overload and to overcome roadblocks to infor-
mation-sharing. The development process of new technology for the battlefield is
too often cumbersome and not related to the needs on the ground. This is partic-
ularly true for military units deployed on nontraditional missions. Because these
units are still such a small part of the military, there has been little support for cre-
ating software programs that improve the process of collecting and sharing data
that are useful for economic development and nation-building efforts. 然而,
the need for this sort of information is growing and becoming increasingly impor-
tant in warfare and economic development.
美国. military is quite skilled in traditional warfare, also called kinetic or
regular warfare. 但, as we have seen in Afghanistan, these traditional approaches
to warfare are far less effective when U.S. forces find themselves in the midst of an
insurgency where the enemy consists of small, decentralized, mobile groups of
insurgents. Counterinsurgency operations have become the norm in Afghanistan,
and a large part of this counterinsurgency involves U.S. troops carrying out eco-
nomic development and nation-building projects. Young soldiers trained for com-
bat have found themselves assigned to such duties as rebuilding economic infra-
structure and serving as liaisons to local government councils. These are tasks for
which they have little training and for which it is difficult to find information on
how to do them most effectively.
Because agriculture is the dominant industry in Afghanistan, 美国. 军队
sent special National Guard units called Agribusiness Development Teams (ADTs)
into Afghanistan with the specific mission of rebuilding that nation’s agriculture
行业. ADTs are company-size units that include both soldiers who are experts
in agriculture and soldiers who are able to provide security for the team. The ADTs
are modeled on the National Guard’s State Partnership Program (SPP), which has
been in operation for two decades. Under the SPP, state National Guards develop
partnerships with “sister” countries and provide training and assistance, 包括
development assistance. An added benefit of SPP teams is the positive relationship
built between U.S. military personnel and civilian and military personnel of the
host nation.
ADTs seem to be an effective model for the U.S. military units that provide
development and nation-building assistance in highly unstable and underdevel-
oped nations. But ADT commanders have made it clear that the model has knowl-
edge management challenges that must be overcome. 例如, these units
92
创新 / Unleashing Ideas
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
The Agility Imperative
have been hampered by the inability to share information with their civilian coun-
terparts on development efforts. Even though information on ADT projects is
mostly unclassified, it is put into classified military systems; once there it cannot
be shared, even with civilian personnel of U.S. government agencies, 例如
我们. Department of Agriculture and USAID. It most certainly cannot be shared
with Afghan government officials or international nongovernmental organizations
on the ground, and this lack of information-sharing makes the coordination of
civil and military efforts nearly impossible. Such coordination is vital in enabling
the military to eventually transfer development efforts to the civilian sector.
Information on the projects and efforts of military teams engaged in develop-
ment efforts is also not easily shared across military units, or from an outgoing
ADT to its replacement ADT. This hinders consistency in development efforts. 它
also means that the lessons learned by one ADT in terms of which strategies are
most effective on the ground are lost, rather than being shared with other military
units. 此外, valuable information collected by ADTs and units on similar
missions is not shared with researchers, which prevents them from conducting
analyses that could lead to better strategies for these teams.
The data collection and management software programs developed for ADTs
will help resolve the problems described above. But because ADTs are already on
the ground and need this information support now, the normal defense depart-
ment process of developing and fielding new technologies will likely be of little
help to ADTs. That process can take years to put new technologies into the hands
of soldiers; 而且, most current information technologies in the military are
aimed at units on traditional military missions. This essentially means that most
information technologies focus on the collection and management of classified
data without concern for the need to share information with civilian counterparts.
Information technologies for ADTs must be developed quickly and placed in
the hands of soldiers as rapidly as possible. These software programs also must be
flexible so they can easily be modified to meet the varying needs of development-
type missions. Developers therefore must work closely with ADT personnel to
ensure that what they create actually meets the needs of the ADTs.
In the traditional technology development process, technology developers,
whether civilian or military, usually design systems to address what they see as a
问题, not necessarily what soldiers have told them is a problem. 开发商
then take their new creation and “toss it over the wall” to soldiers, who often toss
it to the side because it was designed for the wrong problem or it addresses the
right problem but in the wrong way. 而且, because the development-to-
deployment process is so long, when a new technology reaches soldiers, it often
turns out that the problem it was designed to address no longer exists or has
改变了.
A more effective approach, which was used in the case described in this arti-
克莱, is agile development. ADTs face the challenge of operating in rapidly changing
environments and therefore need timely and relevant information for mission
规划. Agile development, as described in the 2001 Manifesto for Agile Software
创新 / 体积 7, 数字 1
93
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart Peintner
发展, is essentially a process that emphasizes interaction between the
developer and the user.2 Developers remain open to adapting a design based on
user feedback. This process should be rapid: users provide feedback, then design-
ers make quick adjustments and return the adjusted product to users for addition-
al feedback. This process can reduce the time between development and deploy-
ment of new technologies from several years to one year or even, as in the case
studies discussed below, to less than a year. Because users are part of this develop-
ment process, the technology they receive meets their needs more effectively. 如果
users’ needs change, the developers quickly modify the technology to meet that
new need.
With funding from the Defense Advanced Research Projects Agency and the
office of the secretary of defense, and under the guidance of the ADT Project at the
Network Science Center at West Point and the Communications-Electronics
Research Development Engineering Center, two technologies were developed for
specific use by units on ADT-type reconstruction missions. Both technologies
were developed with constant interaction and feedback from ADT users, and both
were in the field within the first year of development, a rapidity made possible by
leveraging previous technologies and responding rapidly to users’ needs.
These two technologies—Mobile Task Assistant and XCapture—help soldiers
collect data about their missions. In the case of ADTs, this includes collecting
information on development projects and training programs conducted to help
Afghan farmers. These technologies make it easier for soldiers to share this infor-
mation with other military units, and with civilian government officials and non-
governmental organizations. This information-sharing allows lessons learned
about how to approach stability and how to conduct reconstruction/development
operations to be passed from one military unit to another. It also improves
civil/military coordination of reconstruction efforts. Sharing lessons learned and
removing information barriers to civil/military coordination will go far in helping
unstable and underdeveloped regions get back on their feet.
NATIONAL GUARD AGRIBUSINESS DEVELOPMENT TEAM
KNOWLEDGE MANAGEMENT SYSTEM: A CASE STUDY OF AGILE
SUPPORT FOR NON-KINETIC EFFECTS
One of the emerging challenges in military mission command is the ability to use
the available data and information across missions and issue areas to support com-
manders’ decisionmaking. The “data-to-decision” paradigm refers to this ability to
use data and information to make decisions. The current paradigm is not perfect,
but it is effective when applied to most traditional war fighting conditions. 这是
because the user community and the subject-matter experts understand what crit-
ical information is needed to support their missions and their workflow. 他们还
understand how these information requirements serve as either lagging or leading
indicators of how successfully they are achieving their objectives. 在某些情况下,
94
创新 / Unleashing Ideas
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
The Agility Imperative
this information might include knowledge of how these indicators tie to outcomes
within a particular mission or campaign plan when it is a kinetic (战术上的) 使命.
The picture is not as clear on the non-kinetic side of tactical engagements. 在
most areas that fall outside the tactical intelligence domain, which includes people
and infrastructure, 美国. military has only a small knowledge base. The military
needs a wider range of information in order to understand and address the com-
plex interactions and relationships that occur among the entities in the political,
军队, 经济的, 社会的, 基础设施, and information domain. One reason for
this is that irregular warfare, meaning such things as counterinsurgency and sta-
bility operations, has forced the U.S. Army to support missions such as economic
development and nation-building, which traditionally would be performed by
government agencies like the state department or USAID. No civilian agency could
hope to safely undertake reconstruction work in Afghanistan, 例如, 因为
of the unstable conditions. Although tasked with this reconstruction mission, 这
military has neither the information nor the tools to carry out critical tasks.
This case study addresses how this challenge might be overcome. It describes
the application of the agile systems engineering approach to the development of
knowledge management/information systems designed for use by ADTs in
阿富汗. This case shows the benefits of this innovative approach in develop-
ing and applying technologies that can help military teams involved in non-kinet-
ic missions. Much has been learned in this particular case about how to apply an
agile development approach most effectively.
AGILE SYSTEMS ENGINEERING
Agile software development is one aspect of the larger agile systems engineering
过程. The Manifesto for Agile Software Development details the guidelines used
by agile software developers, and focuses on the following values:
(西德:129) Individuals and interactions over processes and tools
(西德:129) Working software over comprehensive documentation
(西德:129) Customer collaboration over contract negotiation
(西德:129) Responding to change over following a plan
These values apply directly to any effort in which the primary objective is to
produce technology rapidly for a specific customer. The process works when the
customer—in this case the ADT—is dedicated, knowledgeable, and can prioritize
the functions that have the most value. 而且, it results in architecture
designed to meet up-to-date requirements in an environment where the systems
are largely emerging and can change rapidly. These values lead to the 12 原则
that guide agile development:
(西德:129) Our highest priority is to satisfy the customer through early and continuous
delivery of valuable software.
(西德:129) We welcome changing requirements, even late in development. Agile processes
harness change to the customer’s competitive advantage.
创新 / 体积 7, 数字 1
95
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart Peintner
(西德:129) We rapidly deliver working software from a couple of weeks to a couple of
月, with a preference to the shorter timescale.
(西德:129) Business people and developers work together daily throughout the project.
(西德:129) We build projects around motivated individuals, giving them the environment
and support they need and trusting them to get the job done.
(西德:129) The most efficient and effective method of conveying information to and with-
in a development team is face-to-face conversation.
(西德:129) Working software is the primary measure of progress.
(西德:129) Agile processes promote sustainable development. The sponsors, developers,
and users should be able to maintain a rapid pace indefinitely.
(西德:129) Continuous attention to technical excellence and good design enhances agility.
(西德:129) 简单, the art of maximizing the amount of work not done, 是必不可少的.
(西德:129) The best architecture, 要求, and designs emerge from self-organizing
团队.
(西德:129) At regular intervals, the team reflects on how to become more effective, 然后
adjusts its behavior accordingly.
Applying the agile manifesto to systems engineering means increasing the
speed with which new products and systems are designed and delivered to the cus-
tomer. It also entails addressing the uncertainty in future user needs, 操作
状况, and functional requirements.3 Developing information technologies to
meet the needs of ADTs was an ideal opportunity to test this innovative approach
to technology development and deployment.
The development-focused mission of ADTs is different from traditional soft-
ware development processes found in the agile manifesto, in that ADTs primarily
collect unclassified information on their reconstruction efforts, and they need to
share that information with other military units and civilian counterparts that also
are working to rebuild Afghanistan. Most military information technologies are
not designed to deal with unclassified information, and definitely are not designed
to share information. As mentioned above, two technologies were developed
specifically for the ADT Project. One is XCapture, an automated after action
review technology created by AlphaTRAC, 公司, that gathers information on mis-
西翁, including lessons learned. The other is Mobile Task Assistant (mTA), 开发-
oped by SRI International, which focuses on collecting and sharing information on
projects undertaken by ADTs. These two technologies will make it easier for ADTs
to share information, and the information collected will be used to create a rich
database on ADT missions. This database will make it possible for researchers to
analyze and learn from the ADT missions, and thus to improve future strategies for
military teams involved in stability and reconstruction efforts. Application of the
agile development process to the development of XCapture and mTA is described
以下.
96
创新 / Unleashing Ideas
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
The Agility Imperative
XCapture
XCapture is an automated after action review technology that is designed to
improve the information collected about military missions in order to improve
future missions. This automatic system for creating reports gathers and compares
what was supposed to happen on a mission and report what really occurred. 这
system asks what lessons were learned, and then archives the information in a way
that makes it useful for the development of artificial intelligence, 训练, 和
operational support. It is designed to create reports automatically and in various
formats.
The first step in the development of XCapture was to build on an existing tech-
nology known as AlphaACT, which was the result of a Defense Advanced Research
Projects Agency SBIR project conducted by AlphaTRAC. It is web-based software
that goes beyond data management to support and improve the actual decision-
making process. The system uses artificial intelligence to capture and use the expe-
riences of successful decisionmakers. AlphaACT also creates a communitywide
knowledge base, which increases the sharing of knowledge and experiences among
decisionmakers. AlphaACT is currently being used for emergency response train-
ing and is being adapted for small unit leader decisionmaking training.
Experience in commercial software development over the last 15 years has
shown that an adaptive approach helps ensure both project and end-user satisfac-
的. The agile development methodology has been shown to be most effective in
this environment. In the agile development process, “user stories” first establish the
needs and requirements of product users and ultimately drive product develop-
ment and modifications. They do so through what is termed a “scrum process,”
which involves a series of cycles (called sprints) that last two to four weeks, 期间
which a number of user stories are selected for development and testing. 这
process allows the development team to respond quickly to emerging and evolving
product needs and to address bugs quickly and methodically.
Each sprint involves a team working through a full software development
循环. The cycle includes planning, analysis of requirements, 设计, 编码, 单元
testing, and acceptance testing, the latter being when a working product is demon-
strated to stakeholders. The agile development process minimizes overall risk by
identifying and adapting to evolving user needs and allows the project to adapt to
changes quickly. Team composition for the XCapture development project is cross-
functional and self-organizing. Team members take responsibility for tasks that
deliver the functionality a sprint requires.
The XCapture team includes multiple customer representatives, 例如
National Guard soldiers and representatives of the Agribusiness Development
Team program who make a personal commitment to be available to developers
whenever problems or questions arise throughout the process. At the end of each
sprint, the customer representatives review progress and reevaluate priorities.
The XCapture development team collaborates with the full ADT unit in two
方法. 第一的, intensive training, testing, and feedback sessions take place during
创新 / 体积 7, 数字 1
97
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart Peintner
数字 1. The Agile Product Development Process
workshops that take place periodically at ADT home station or mobilization sites.
第二, soldiers at the home station and in the field conduct ongoing application
testing, all the while providing feedback through email and internal social forums.
The XCapture development process begins with a sprint kickoff session. 这
tasks to be accomplished in each sprint are based on the stakeholders’ needs as they
are defined at the beginning of the sprint. These needs are expressed as user sto-
里斯, which essentially serve as user requirements during the agile development
过程. Stakeholders include the intended XCapture end users and other interest-
ed parties, such as leaders, managers, security personnel, infrastructure managers,
third-party users, system administrators, and even the development team itself.
During the kickoff session, the team chooses user stories that will be addressed
during the sprint and assigns a task leader and task team for each user story.
Addressing some user stories may require more than a single sprint. These stories
are termed “epics” and are broken into sub-units, each of which can be addressed
in a single sprint.
The team meets daily during the sprint. The customer representative and inter-
ested stakeholders can observe at these meetings, during which team members
report what they did the previous day, what they intend to do today, and what their
roadblocks are—a process that exposes problems as they arise.
Each sprint culminates in a sprint review that features the product modifica-
tions or extensions for each user story. Unit and integrated testing data and results
98
创新 / Unleashing Ideas
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
The Agility Imperative
数字 2. Mobile Task Assistant
are also presented. The customer representatives review each modification or
extension and declare whether or not the user’s need was met to their satisfaction.
The user story is then either retired or goes into further development. 开发-
opment cycle continues with a kickoff for the next sprint.
In addition to the testing done within each sprint, the entire team breaks from
the development process to test the system thoroughly after each major release of
product improvements. The agile development process as applied to the XCapture
project is illustrated in Figure 1 (previous page).
桌子 1 (following page) presents an illustrative sample of the initial user sto-
ries that were generated to begin the XCapture development project in June 2011.
MOBILE TASK ASSISTANT
Mobile Task Assistant (mTA) supports ADTs in developing and managing agri-
business projects by facilitating data collection, data sharing, and report genera-
的. This Android-based mTA app was developed over the course of eight months
using the agile process described above. mTA is the mobile companion to the (pre-
创新 / 体积 7, 数字 1
99
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart Peintner
桌子 1. Sample of initial user stories that were generated to begin the XCapture
development project in June 2011
existing) Task Assistant system, a web-based collaborative task management tech-
nology that helps teams coordinate efforts, gather and share data, and disseminate
knowledge and procedures.
Along with its web-based counterpart, mTA helps ADT team members capture
笔记, 图片, audio recordings, and GPS locations more efficiently, and then to
organize that information and make it accessible to both the current team and
future units. At the base, soldiers develop forms and instructional guidelines using
the web-based interface. The forms and guidelines are then downloaded onto the
100
创新 / Unleashing Ideas
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
The Agility Imperative
数字 3. Web-based Task Assistant view of a mission tasklist for the ADTs
Android devices for use in the field. When returning from the field, soldiers fill out
forms and provide images, and sound recordings. These items are merged into the
shared Task Assistant database, which is then used to generate reports and support
future field missions.
Need for agile development
The compressed timeline of less than one year to deploy this new technology
necessitated the agile development approach. The classical approach of gathering
要求, developing the software, then throwing it over the wall was infeasi-
ble for three main reasons. 第一的, gathering requirements for the software required
interaction with several National Guard units, some of which were deployed and
therefore initially unavailable, some of which were too inexperienced at the start of
the project to give valuable feedback, and others that had intense pre-deployment
training schedules that limited our access.
第二, the utility of the software depended primarily on how easily it fit into
the workflow of the mobile teams. ADT teams cannot afford to spend their time
interacting with the software and device—the technology had to perform its func-
创新 / 体积 7, 数字 1
101
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart Peintner
tions without compromising the personal relationships between the team and local
residents. 所以, the design and development process had to center on elicit-
ing design-enabling knowledge from the target users.
最后, the people driving the project had not fully understood the concept of
operations for the software and device. 可以说, this is true for nearly all “ground-
up” software development, which is a primary cause of failure for the waterfall
method of product development. The mTA app represented a substantial change
in team operations, in that its effects and primary values could not be anticipated,
even by experienced domain experts. 此外, just as technology experts do
not initially understand the domain, domain experts do not initially understand
what is possible in software development. 因此, it is difficult to develop a prod-
uct vision that fits into both time and resource constraints, and the vision must
adapt over time. Seeing interim versions of the product shows the domain expert
what is possible and helps generate new ideas.
For the reasons just mentioned, the development and design processes neces-
sarily began before the first “requirements gathering” user event, 因此
well before developers had a complete picture of the eventual uses for the software.
The lack of a complete picture had a strong effect on the order and priorities of the
initial development efforts. The team’s initial development efforts focused on the
underlying client-server interactions and the internal structure of the app. The lack
of a fully specified design actually promoted an agile-friendly design environment;
the infrastructure that was first developed supported a wide range of layouts and
navigation strategies.
Based on experiences on other Personal Assistant that Learns or PAL pro-
克, developers had been working successfully with users to define the concept
of operations in their domains. Initial meetings were held with the ADT members
to understand their roles and to begin to develop the concept of operations.
Developers had a series of meetings with ADT members returning from theater,
ADT members preparing for their second deployments, and ADT members who
were just beginning to train for their first deployment.
Tailoring the standard agile methodology to the situation
While the initial infrastructure was being developed, design efforts focused on
mocking up several ideas to present at the first meeting with the user. The first user
meeting did not require a prototype, but developers believed that presenting a vari-
ety of highly realistic mockups, along with demos of the existing web-based tool,
would elicit more valuable information than a single real but highly limited exam-
ple of working code. This first meeting gave developers the input they needed to
commit to a first design of the layout and navigation.
Developers then tailored the standard agile process in one other significant
方式: they did not hold the typical 15-minute weekly meetings, as Task Assistant
provides collaboration and task management services that serve the same function
as these meetings. Using the tool, developers efficiently tracked who was doing
102
创新 / Unleashing Ideas
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
The Agility Imperative
桌子 2. ADT User Events
什么, gathered ideas for improvements, reviewed requirements elicited at the last
user event, defined short-term goals, and reported on the results of testing efforts.
因此, almost no time was spent in meetings coordinating, and meeting time
was used instead to discuss user feedback, brainstorm solutions, and argue for dif-
ferent technical approaches—all of which moved the product forward quickly.
Iterations and ADT user events
Developers were fortunate to go through several iterations with the ADT users—a
surprising number in fact, given the schedules and training demands of recently
deployed and soon-to-be-deployed soldiers. Each of the iterations consisted of a
user event, a goal-setting session, and a development sprint.
桌子 2 outlines the nature of each user event and the level of prototype avail-
able for each. After each event, developers updated the design, reprioritized fea-
特雷斯, and set goals for the next release. Several factors influenced which improve-
ments to make between events:
(西德:129) Level of uncertainty. Uncertainty in the value or design of a feature is cause for
inclusion in the next user event, given that it may be necessary to maximize the
number of remaining iterations to get the feature right.
(西德:129) Difficulty of implementation. Complex elements of the product require more
testing and user time to get issues ironed out.
(西德:129) Visibility of problem or solution. When users perceive improvements in the
product from session to session, and when they understand that their feedback
finds expression in the product, their feedback improves in quantity and level of
insight. 反过来, an unaddressed deficiency that users perceive as major
reduces their faith in the process and the development team.
The natural progression of the user events tends toward an increasingly capa-
ble product, an increasingly refined concept of operations, and a general slowing
of the rate of change in each. In the case of mTA, one or two additional iterations
would have significantly improved the initial delivery. These iterations have con-
创新 / 体积 7, 数字 1
103
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart Peintner
tinued post-deployment, as developers continue to improve the system based on
feedback from the field.
结果
This agile development process quickly produced a working prototype. 虽然
technical and usability issues remained at the time of delivery, initial reports from
the field, which outlined the positives and negatives of their experience with mTA,
showed clearly that the ADT members were pleased with the development process
and were eager to continue it. Developers have continued to produce iterations,
sending new releases to the field every few months.
The agile process uncovered several non-software issues that had to be
addressed to enable effective use. A lot of time was spent exploring such things as
gloves, rugged cases, and input devices—for example, keyboard, stylus, and pouch-
es—that had to be tried by a number of soldiers before something workable was
成立.
Other issues stemmed from limitations in the field, such as the lack of wireless
连接性, the need for and the challenge in creating hard cases for the device,
and the wide variety of lighting conditions. Early interaction with the users forced
developers to tackle these issues early enough to be able to design a solution into
the final product. The lack of wireless connectivity on base, 例如, added a
significant technical hurdle that required several months to resolve.
The final version of the app differs significantly from the initial design. 一些
features originally marked as “must have” did not make it into the system because
no user session ever showed the need for them. Implementing those features
would have taken time away from higher priority tasks and cluttered the interface,
making it harder to learn to use. 此外, features that were secondary and
somewhat hidden in the web interface (例如, displaying a location field) ended up
being prominent, easily created features of mTA. All of these differences point to
the value of the iterative, agile process. Using the classic development process, 很少
challenges would have been identified or solved by the time users required.
结论
在过去的十年, we have overcome many planning, 执行, and informa-
tional challenges that we faced when conducting both stability and response-and-
recovery operations, but many challenges remain in terms of improving collabora-
tion between military and civilian government agencies, and between those agen-
cies and NGOs and the private sector. Particularly challenging in both stability
operations and response-and-recovery operations is the impact of our inability to
share information in real time, or close to real time. We face policy, procedural,
and technology barriers, in addition to the normally degraded infrastructure that
accompanies both situations. The ability to address the barriers and challenges
faced in these complex situations will require our continued efforts to reduce the
104
创新 / Unleashing Ideas
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
The Agility Imperative
time it takes to produce solutions by applying the agile development process to
more than just technology solutions.
Traditional technology development followed a classical format that involved
creating requirements—often without input from end users—building technology
to meet those requirements, tossing the new technology “over the wall” to soldiers,
and then leaving them to “like it or lump it.” The typical result has been the devel-
opment of technologies that users neither liked nor needed. 或者, more accurately,
the development of technologies no longer needed because the situation had
changed long before the technology ever made it from drawing room to the field.
The agile development process takes a different view, which essentially is that
the end user—in this case the warfighter—is all that matters, not what the technol-
ogy developers think they want or should want. The agile approach requires con-
stant interaction with users during the development process, and making constant
adjustments as the needs of the user change. As this case study of XCapture and
mTA show, when effectively applied, the agile development approach can lead not
only to faster deployment of new technologies but also to the creation of technolo-
gies that meet users’ needs more effectively, are viewed more favorably, 在那里-
fore are used more widely by warfighters.
The success of this approach in developing technologies that actually provide
benefits to soldiers on ADT reconstruction missions provides lessons that leaders
in both military organizations and civilian agencies should learn. Far too often,
managers at the top of organizations have no real understanding of the situation
on the ground or of the needs of their personnel who are out there conducting
运营, whether they be stability operations or disaster response and recovery.
Information-sharing, both across organizations and among the levels within an
组织, is critical for coordinating efforts.
What is needed, 然后, is a new approach that supports the development of
technologies and the adoption of policies that can improve information-sharing
and coordination. This approach must force managers at the top to allow those on
the ground a greater voice in decisions, not only regarding the development of
technologies that meet their needs but in changing policies in ways that can
improve their ability to share information and coordinate with others on the
地面. This case study shows how quickly needs can be met when leaders at the
top of organizations, end users on the ground, and developers engage in constant
dialogue in which each listens to and understands the needs of the others.
1. Kilcullen, 大卫. The Accidental Guerilla: Fighting Small Wars in the Midst of a Big One. 牛津,
England: 牛津大学出版社, 2009.
2. Manifesto for Agile Software Development, http://agilemanifesto.org. See also Lingfeng Wang,
“Agility counts in developing small-size software,” Potentials, IEEE, 卷. 26, 不. 6 (Nov.-Dec. 2007):
16-23.
3. D. Oxenham. “Agile approaches to meet complex system of system engineering challenges: A
Defence Perspective,” System of Systems Engineering (SoSE), 2010 5th International Conference,
PP. 1-6, 22-24, 六月 2010. V. Rahimian, and R. Ramsin, “Designing an agile methodology for
创新 / 体积 7, 数字 1
105
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
Eric Peck, Lynndee Kemmet, Ray McGowan, C. Reed Hodgin, and Bart Peintner
mobile software development: A hybrid method engineering approach,” Research Challenges in
Information Science, 2008. Second International Conference, 六月 2008; A.M. Sen and K.
Hemachandran, “Elicitation of Goals in Requirements Engineering Using Agile Methods,”
Computer Software and Applications Conference Workshops, 2010 IEEE 34th Annual, 七月 2010.
106
创新 / Unleashing Ideas
从http下载的://direct.mit.edu/itgg/article-pdf/7/1/91/704889/inov_a_00118.pdf by guest on 08 九月 2023
