Air Force ADP: Depot Maintenance System Development Risks Are High

Published by the Government Accountability Office on 1990-05-25.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

                 United   States   General   Accounting   Office

GAO              Report to the Chairman, Subcommittee
                 on Readiness, Committee on Armed
                 Services, House of Representatives

May 1996
                 AIR FORCE ADP
                 Depot Maintenance
                 System Development
                 Risks’Are High

                   United States
                   General Accounting  Office
                   Washington, D.C. 20548

                   Information      Management        and
                   Technology      Division


                   May 25,199O

                   The Honorable Earl Hutto
                   Chairman, Subcommittee on Readiness
                   Committee on Armed Services
                   House of Representatives

                   Dear Mr. Chairman:

                   In response to your concerns about development problems, rising cost,
                   and schedule delays, we reviewed the Air Force Logistics Command’s
                   attempt to develop the Depot Maintenance Management Information
                   System (DMMIS). This new automated system is intended to modernize
                   and improve data processing support at the Command’s six maintenance
                   centers. In 1984, the Command estimated the system would cost about
                   $85 million and be fully operational in February 1989. The Command
                   now estimates it will cost $242.4 million and be fully operational in Sep-
                   tember 1993.

                   To develop DMMIS, the Command plans to purchase hardware and adapt
                   commercial, off-the-shelf software called Manufacturing Resource Plan-
                   ning (MRP II)’ for each of 19 product divisions at its six centers. Our spe-
                   cific objectives were to determine if (1) the adaptation of MRP II software
                   will meet the needs of the Command’s depot operations, (2) the acquisi-
                   tion cost estimates are reasonable, and (3) the development schedule can
                   be met. Appendix I provides details on our objectives, scope, and meth-
                   odology. Appendix II provides additional information on Air Force depot
                   maintenance operations and the 19 product divisions at the six depot
                   maintenance centers.

                         software will be more difficult to adapt to the depot environment
Results in Brief   MRP II
                   than the Command had expected. Originally expected to meet 90-95 per-
                   cent of DMMIS requirements, the Command currently estimates that MRP II
                   will meet only 65 to 70 percent of the system requirements-however,
                   one contractor study concluded that 51 percent is more realistic. This
                   means that the Command must significantly change the MRP II software,
                   which increases the risks of not staying on schedule and within budget.

                    ‘MRP II IS a widely accepted industry approach to integrated manufacturing planrung and control. It
                   includes software. procedures. and controls to coordinate. manage, and operate a complex manufac-
                   tunng or repair facility by following a single master production plan.

                   Page 1                                   GAO/IMTJZSO46       Air Force Depot Maintenance     System

Further adding to risks is the amount of work that could be needed to
prepare depots for DMMIS. At one location, over 4 years and 184,000
staff hours were needed primarily to clean up the existing data bases
that DMMIS will use, and the Air Force expects a similar effort at each of
the other five locations.

In June 1989, the Command recognized that without a change in devel-
opment strategy, DMMIS could not be completed for $242.4 million and by
September 1993. As a result, the Command is currently negotiating
changes with Grumman Data Systems, Incorporated, the primary con-
tractor, that will significantly alter the DMMIS scope and implementation
approach. Under the revised approach, the Command no longer plans to
complete a prototype at 1 of the 19 product divisions before beginning
work at other divisions. Instead it will begin adapting the MRP II software
at five product divisions simultaneously.

Because of the significant modifications to MRP II software that are nec-
essary, the new approach adds risk to an already risky program. By
waiting until February 1991-the scheduled completion date of the pro-
totype-the    Command would know whether the MRP II concept will meet
its needs for the DMMIS program. Considering the relatively short time
until the prototype is to be completed, we believe the Command should
complete the DMMIS prototype before beginning development efforts at
other product divisions.

depot-level maintenance to keep Air Force units and weapons systems in
a state of readiness. The Command relies on computer technology to
provide the enormous amount of information needed to accomplish its
mission. Many of the Command’s computer systems originated in the
1950s and 1960s and, like other systems that date back to this era, have
not kept pace with advances in computer technology.

In 1984, the Command began the DMMIS program to improve the overall
efficiency and effectiveness of its depot maintenance operations. It
expects DMMIS to provide repair depots with on-line capability to fore-
cast work loads; schedule repair activities; track and control inventories;
program manpower, materials, and other resources; and track and man-
age production costs, Currently, these activities are done either manu-
ally or by more than 50 existing systems, 29 of which the Command
plans to replace with DMMIS. (To do all this, the DMMIS system will include
about 2.5 million lines of code.)

 Page 2                        GAO/IMTJZG!#O46   Air Force Depot Maintenance   System
~-   .--

                   This is an enormous undertaking considering the Command plans to
                   implement DMMIS at all five of its Air Logistics Centers and the Aero-
                   space Guidance and Metrology Center. The Air Force Logistics Com-
                   mand operates a depot maintenance industrial complex with annual
                   work valued at $2.5 billion. This maintenance community employs over
                   39,000 persons, and has invested $4.5 billion in facilities and equipment.
                   Air Force depot maintenance is responsible for the repair and rebuilding
                   of 1,200 aircraft, 1,200 missiles, 6,400 engines, and 1.1 million other
                   repairable items. Compounding the number of items for which the
                   depots are responsible is the fact that many repairs are extremely labor
                   intensive. For example, the repair of one aircraft, such as the B-52, can
                   require 50,000 hours. Annually, over 44 million hours are expended by
                   the depots to perform maintenance.

                   The depot maintenance process can be very complex. For example, a sin-
                   gle engine can comprise 1,800 parts or subcomponents, require 12
                   departments to become involved in its maintenance, spread out over five
                   to six subassembly areas before the engine parts are sent to a final
                   assembly area. Periodically, each engine is completely dismantled and
                   every part examined, See appendix II for further information on the
                   depot maintenance process.

                   The original DMMIS acquisition strategy called for development of three
                   distinct systems-production,    financial, and resource management.
                   Because of technological difficulties and high risks in integrating these
                   three systems, the Command abandoned this strategy in 1985, and later
                   restructured the DMMIS development strategy into a single system design
                   using commercial MRP II software. Recognizing the complexities of the
                   proposed system and the depot maintenance environment, the Command
                   plans to develop DMMIS in three phases.

           l Phase I involved developing an interim system called the Exchangeable
             Production System to improve controls over new parts managed at the
             Command’s Air Logistic Centers.
           m Phase II called for the completion of a system prototype at the Ogden
             Air Logistics Center Industrial Products and Landing Gear Division, 1 of
              19 product divisions scheduled for DMMIS automation. The Command
             intended to test and evaluate DMMIS in an operational environment
             before implementing it at the other product divisions.
             Phase III involved full-scale development and system implementation.

                   Page 3                        GAO/lMTEG9046   Air Force Depot Maintenance   System

                       The Command completed Phase I implementation at all five Air Logis-
                       tics Centers in July 1988 at an estimated cost of $28.1 million. In Janu-
                       ary 1988, the Command awarded an $84 million contract to Grumman
                       Data Systems to design, develop, implement, and maintain phases II and
                       III of DMMIS. Other program costs-totaling   over $130 million-will
                       include (1) software and hardware maintenance for phase I, (2) com-
                       puter terminals and other hardware used by maintenance workers on
                       the shop floor, (3) contractor technical support to help the Command
                       manage the program and oversee system testing, and (4) several other
                       miscellaneous items.

                       According to the Command, successful completion of the DMMIS program
                       could save as much as $706 million over the 8-year useful life of the
                       system. Most of these benefits are attributable to improvements in pro-
                       ductivity and reduced inventory costs.

DMMIS Is a High Risk   government, consider DMMIS a very complex and high-risk undertaking.
Program                The Command’s ability to adapt MRP II software to meet DMMIS require-
                       ments is clearly the most critical issue yet to be resolved. Other issues
                       affecting the program’s success include the size and complexity of the
                       program development effort and the lack of management continuity.

MRP II May Not Meet    As discussed earlier, in designing the DMMIS development strategy in
                       1984, the Command intended to develop three separate systems which
DMMIS Requirements     would interface with one another. When this proved technically infeasi-
                       ble, the Command hired Deloitte Haskins & Sells to evaluate alternatives
                       and recommend a workable solution. In its May 1985 report, Deloitte
                       recommended that the Command use a commercially available software
                       package-called MRP n-in lieu of developing a completely new system.
                       Deloitte estimated MRP II would satisfy from 90-95 percent of DMMIS
                       requirements. A subsequent Air Force Audit Agency assessment of
                       Deloitte’s projections found that more realistically, MRP II would satisfy
                       about 65-70 percent of DMMIS requirements. In January 1988, the Com-
                       mand awarded the DMMIS development contract to Grumman Data Sys-
                       tems, Incorporated, expecting that 65-70 percent of the requirements
                       would be met by MRP II.

                        In May 1988, Entek, Inc., (a contractor hired by the Command to pro-
                        vide technical support) reported that the MRP II software would meet
                        only 5 1 percent of DMMIS requirements. Consequently, Entek concluded

                        Page 4                        GAO/IhlTEG90-46   Air Force Depot Maintenance   System

                            that more extensive modifications-as  well as a significant amount of
                            newly developed software-were   necessary to meet DMMIS requirements.
                            Although these would probably raise costs and delay the schedule,
                            neither the Command nor Grumman tried to make new estimates.

                            About the same time, Synergy, Incorporated, (another independent con-
                            tractor) assessed the potential operational impacts of DMMIS on aircraft
                            readiness. Synergy concluded that DMMIS could, if properly developed
                            and implemented, increase the number of aircraft available during both
                            peace and wartime situations. Synergy raised concerns, however, about
                            adapting MRP II to the complex depot repair environment: MRP II had
                            worked in manufacturing processes (where the tasks are more predict-
                            able and routine), but adapting it to the depot repair environment was
                            risky. Synergy had surveyed several private sector companies and
                            found limited success in modifying MRP II to repair or overhaul-type
                            operations. However, the Command believed adapting MRP II would be
                            easier than starting from scratch.

Additional Risks Could      The time and effort needed to prepare each location for DMMIS adds to
                            the risk that the program may not be completed on schedule. Site prepa-
Affect, Cost and Schedule   ration activities? for the first implementation site in Ogden took 4 years
                            and 184,000 hours to complete. While it is difficult to judge how long it
                            will take to prepare the other locations, the Command expects a similar
                            level of effort will be needed. Another concern was raised by the Air
                            Force’s Cost Analysis Improvement Group, which concluded that to
                            meet the September 1993 target completion date, DMMIS would likely
                            need an additional $32.4 million to cover site-unique requirements not
                            previously identified.

                            The Secretary of Defense’s Major Automated Information System
                            Review Committee found that DMMIS risk management needed strength-
                            ening. In June 1989, this committee directed the Command to develop a
                            plan to identify and address DMMIS developmental risks. According to
                            that plan, using MRP II software is a high-risk issue involving cost and
                            schedule estimates and the success of DMMIS itself.

                            The Command originally underestimated DMMIS complexity. The Vice
                            Commander for the Logistics Management Systems Center told us that

                            ‘Site preparation activities include taking a physical mventorv of all parts and materials, cleaning up
                            existing data bases that will be integrated into DMMIS, gather&g new data that will be entered into
                            DMMIS. and mstalling power lines.

                            Page 5                                    GAO/IMTEG904         Air Force Depot Maintenance      System

                          the DMMIS program has changed so dramatically since its inception in
                          1984 that it no longer resembles the program initially approved by the
                          Air Force. From time to time, the Command reduced the program’s scope
                          to stay within cost and schedule estimates. Among other things, the
                          Command has reduced the number of systems DMMIS will replace from
                          43 to 29 and decided to prototype the system at one rather than two
                          product divisions. These changes as well as the program’s cost and
                          schedule growth prompted the Major Automated System Review Com-
                          mittee to assume oversight of the program in September 1988. Previ-
                          ously, the Air Force’s Automated Information System Acquisition
                          Review Council had that responsibility.

Lack of Management        The lack of management continuity has also increased the program’s
                          development risks. The Command has changed program managers three
Continuity Increases      times since the DMMIS program began in 1984. In addition, the current
DMMIS Development Risks   program manager, who arrived in June 1989, anticipates retiring and
                          leaving the program in June 1991. Without trying to determine why
                          these changes occurred or the specific impact each had on DMMIS'S devel-
                          opment, this rapid turnover of top management is not conducive to pro-
                          gram stability, particularly for a program of this size and complexity.
                          According to a recent General Services Administration study of 18 fed-
                          eral computer modernization programs, the program manager’s ability
                          was the single most important factor in successful system development
                          efforts. While we are not questioning the ability of the various DMMIS
                          program managers, the lack of managerial continuity, in our view,
                          clearly puts DMMIS at risk.

                          In June 1989, the Command revised its development strategy for DMMIS
The Command Has           to stay within cost and schedule estimates. It adopted a Deloitte Haskins
Revised Development       & Sells recommendation to implement a technique-called      Conference
Strategy but Risks        Room Pilot-which     allows the system users to work directly with the
                          system development contractor to identify requirements unique to their
Remain                    working environment, develop software, and resolve potential operating
                          problems. This early and direct involvement of the users, in theory,
                          allows developers to compress the development schedule.

                          Another part of the Command’s revised strategy, however, may actually
                          increase risks. ~Previously, the Command intended to complete installa-
                          tion and testing of DMMIS for one product division-the    Industrial Prod-
                          ucts and Landing Gear Division-at     the Ogden site (currently scheduled
                          for February 1991) before beginning work at additional sites. But now it

                          Page 6                        GAO/IMTJ3G9O46   Air Force Depot Maintenance   System

                  has begun work at two other product divisions at the Ogden site and
                  plans to begin system development and installation at two other loca-
                  tions in July 1990. Therefore, the Command will be developing DMMIS fo
                  five product divisions at three locations at the same time, before the
                  system has been tested and proven to work anywhere. Recognizing the
                  risk entailed by starting work at other product divisions before the
                  Ogden prototype is completed, DMMIS'S program manager believes this
                  strategy is necessary if the Command is to finish within current cost and
                  schedule estimates.

                  Even with this revised structure, however, recent information from the
                  DMMIS  program office shows that the schedule is beginning to slip. Under
                  the restructured approach, for example, development work at the two
                  engine divisions was scheduled to start in February 1990. The latest
                  program office informal estimate is July 1990. Nevertheless, the Com-
                  mand believes it can still meet its overall September 1993 target comple-
                  tion date.

                  Clearly, DMMIS is one of the largest and most complicated development
Conclusions and   efforts undertaken by the Air Force Logistics Command. This in itself is
Recommendations   enough to make the program a high-risk venture. There are three other
                  issues, however, which tend to increase the developmental risks. First,
                  the MRP II software will now require more extensive software changes
                  than originally planned, and until the Command completes the proto-
                  type system in February 1991 I it is uncertain if MRP II will effectively
                  work in the depot environment. Yet, under the Command’s recent pro-
                  gram restructure, it plans to begin adapting the MRP II software at two
                  other product divisions before the prototype system is complete. This,
                  we believe, is adding an unnecessary risk to an already risky program.

                  Second, a huge effort may be needed to prepare each location for DMMIS.
                  Since site preparation activities, which mainly involve cleaning up the
                  data bases, took over 4 years to complete at the first location, a similar
                  effort at the other locations could have an impact on the system

                  Third, since its inception, the DMMIS program has had four program man-
                  agers. In addition, the current program manager, who has been with the
                  program since June 1989, may be leaving next year. This lack of man-
                  agement continuity, in our view, also adds risk to the program.


    To address these issues, we recommend that the Secretary of Defense
    direct the Secretary of the Air Force to:

l   complete the prototype system at the Ogden Industrial Products and
    Landing Gear Division before committing resources to develop DMMIS at
    other product divisions,
l   restructure the DMMIS schedule and cost estimates to reflect this change
    in development strategy, and
l   ensure management continuity for the DMMIS program.

    In accordance with your wishes, we did not obtain official agency com-
    ments on this report. We did, however, discuss its contents with Depart-
    ment of Defense and Air Force officials and have included their
    comments where appropriate. In a recent meeting, the program manager
    told us he concurs in principle with our recommendations and will
    reconsider his position on the concurrent system development at the two
    engine divisions.

    As agreed with your office, unless you publicly announce the report’s
    contents earlier, we plan no further distribution of it until 30 days from
    the date of this letter. At that time, we will provide copies of this report
    to the Secretary of Defense and the Secretary of the Air Force. We will
    also make copies available to other interested parties upon request. This
    work was performed under the direction of Samuel W. Bowlin, Director,
    Defense and Security Information Systems, who can be reached at (202)
    275-4649. Other major contributors are listed in appendix III.

    Sincerely yours,

    Ralph V. Carlone
    Assistant Comptroller General

    Page 8                         GAO/IMTJ3G90-46   Air Force Depot Maintenance   System
Page 9   GAO/lMTEC9046   Air Force Depot Maintenance   System

Appendix I
Objectives, Scope, and
Appendix II                                                                                                 14
An Overview of Depot
Appendix III                                                                                                16
Major Contributors to
This Report
                         Table II. 1: Location of Air Force’s 19 Product Divisions                          14


                         ADP       automated data processing
                         DMMIS     Depot Maintenance Management Information System
                         GAO       General Accounting Office
                         IMTEC     Information Management and Technology Division
                         MRPII     Manufacturing Resource Planning

                         Page 10                        GAO/lMTJ3GSO-46   Air Force Depot Maintenance   System
Page 11   GAO/lMTEG90-46   Air Force Depot Maintenance   System
Appendix I

Objectives, Scope, and Methodology

               Concerns over past system development problems and significant cost
               and schedule growth prompted the Chairman, Subcommittee on Readi-
               ness, House Committee on Armed Services, to ask us to review the sta-
               tus of the Air Force’s Depot Maintenance Management Information
               System (DMMIS). After further discussions with his office, we agreed to
               assess whether the DMMIS development approach will result in a system
               that meets requirements within the expected cost and schedule. Specifi-
               cally, our objectives were to determine whether (1) the commercial MRP II
               software will satisfy DMMIS functional requirements, (2) the acquisition
               cost estimates are reasonable, and (3) the development schedule can be

               To address the first objective, we focused on the (1) feasibility of apply-
                ing commercial MRP II software in a repair environment, and (2) method-
                ology used by the development contractor and the Air Force Audit
                Agency to estimate the expected effectiveness of the MRP II software in
                meeting DMMIS requirements. We obtained and reviewed industry trade
               journals and published periodicals on MRP II software. We analyzed sup-
                porting Command and contractor documentation as well as independent
                assessments. To understand how well MRP II software works in a repair
                environment, we spoke with private industry and Defense officials who
                have worked with or are now implementing an MKP II system. We also
               visited the DMMIS test site located at the Air Logistic Center at Ogden,
                Utah, and spoke with system users on the pros and cons of implement-
                ing MRP II in a repair environment and the level of pre-implementation
                activities required.

               To determine whether the approved DMMIS acquisition cost and schedule
               estimates can reasonably be met, we updated the information presented
               in a prior GAO report] on cost and schedule growth. We obtained and
               analyzed budgetary and contract cost data, reviewed current project sta-
               tus reports and pertinent program documents such as the program office
               estimate, cost/benefit analysis, and program decision papers. We also
               tracked key program milestones and spoke with Command and program
               officials regarding the accuracy and reasonableness of these milestones.

               Our review was conducted from January 1989 to February 1990, prima-
               rily at the DMMIS Program Management Office in Fairborn, Ohio; the
               Directorate of Maintenance at the Air Force Logistics Command at

               ‘Air Force ADP: Logistics Systems Modemizatlon   Costs Continue to Increase IGAO/IMTEC-89-7I3,
               Dec. 28. 1988).

               Page 12                                 GAO/IMTFX-90-46     Air Force Depot Maintenance      System
Appendix I
Objectives, Scope, and Methodology

Wright-Patterson Air Force Base, Ohio; and the            DMMIS   test site at the Air
Logistics Center in Ogden, Utah.

In accordance with the requester’s wishes, we did not obtain official
agency comments on this report. However, we discussed the contents
with Department of Defense and Air Force officials and have included
their comments where appropriate. We performed our work in accor-
dance with generally accepted government auditing standards.

Page 13                              GAO/IMTEG90-46   Air Force Depot Maintenance   System
An Overview of Depot Maintenance

                                            The Air Force Depot Maintenance community consists of 19 product
                                            divisions that perform repair, modifications, and manufacturing in sup-
                                            port of Air Force weapons systems. Each division specializes in a certain
                                            type of work. These product divisions are located at the five Air Logis-
                                            tics Centers] and the Aerospace Guidance and Metrology Center, Ohio.
                                            Each operating site has three product divisions, except for one, which
                                            has four. Each product division is comprised of a Production Branch,
                                            Engineering and Planning Branch, and Scheduling and Inventory Con-
                                            trol Branch which makes them, to a great extent, independent. Each
                                            product division operates independently, and can be considered a sepa-
                                            rate business enterprise.

                                            Product divisions located at the same centers perform work for each
                                            other in a manner similar to that of a manufacturer who sends work out
                                            to a specialty shop. When such work is completed, it is returned to the
                                            originating division for assembly into a finished product. Table II. 1
                                            below shows the 19 different product divisions and the center where
                                            they are located.

Table 11.1:Location   of Air Force’s   19
Product Divisions                           Center                                            Product division
                                            Aerospace Gurdance and Metrology Center            1 Aircraft products
                                                                                               2. Mrssrle products
                                                                                               3. Support equipment
                                            Oklahoma 0ty Air Logtstrcs Center                  4. Aircraft (B-52, E-3)
                                                                                               5. Propulsion
                                                                                               6 Accessories
                                            Ogden Air Logrstrcs Center                         7 Arrcraft (F-16, F-4)
                                                                                               8 Mwle and aircraft systems
                                                                                               9 Industrial products and landtng gear
                                            San Antonio Atr Logisttcs Center                  IO Aircraft (C5A, C-130)
                                                                                              11 Engines
                                                                                              12 Technology repair
                                            Sacramento Arr Logrstlcs Center                   13. Aircraft (A-10, F-l 11)
                                                                                              14 Flight instruments and pneudraulics
                                                                                              15 Communtcatrons and electronics
                                                                                              16. Industrial products and electronrc
                                            Warner Robrns Arr Logrstrcs Center                17 Aircraft (F-15, C-130)
                                                                                              18. Airborne electronics
                                                                                              19 Industrial products

                                            ‘These centers are located at Hill Air Force Base. Utah; Tinker Aw Force I&e. Oklahoma; McClellan
                                            Air Force Base. Califonua: Kelly Air Force Base, Texas; and Robins .4ir Force Base, Georgia.

                                            Page 14                                 GAO/IMTEG9046       Air Force Depot Maintenance    System
Appendix II
An Overview   of Depot Maintenance

There are two basic categories of work at the depots-permanent       and
temporary. Permanent work involves repetitive maintenance require-
ments that are anticipated in advance of actual accomplishment. Having
been anticipated, they are managed on a programmed basis. Permanent
work loads include aircraft, missiles, and engines. In contrast, tempo-
rary work loads have the characteristics of generating spontaneously
and being non-repetitive, such as the manufacturing of new products.

For permanent work loads, the customers provide work load projections
covering several years. Prior to the beginning of each fiscal year, air-
craft, engines, and other items are negotiated for the year. During the
negotiations, there is a need to project the work load against available
resources (i.e., staffing, equipment, facilities, materials). There are sev-
eral work loads that are negotiated and managed on other than a yearly
basis. For example, depot-level aircraft engine maintenance facilities
only exist at San Antonio and Oklahoma City Air Logistics Centers.
Engine work is negotiated on a quarterly basis and includes several
repair concepts. One of these involves conplete engine disassembly,
inspection of all components, routing of repairable components to the
appropriate repair shop, obtaining new or repaired components for
engine assembly, and engine testing. The other concept is on-conditions
maintenance. This is a repair program whereby each engine or engine
module is disassembled only to the extent necessary to make repairs
and/or replace worn, damaged, and life-limited parts.

Page 16                              GAO/IMTEG9046   Air Force Depot Maintenance   System
Appwdix    III

Major Contributors to This Report

                       John B. Stephenson, Assistant Director
Information            Suzanne M. Burns, Assignment Manager
Management and         Sanford F. Reigle, Adviser
Technology Division,
Washington, D.C.

                       Daniel V. Loesch, Regional Management Representative
Cincinnati Regional    Steven M. Hunter, Evaluator-in-Charge
Office                 Barbara L. Centers, Evaluator

(5103il)               Page 16                     GAO/IMTEGSO46   Air Force Depot Maintenance   System
Requests for copies of GAO reports      should be sent to:

U.S. General Accounting       Office
Post Office Box 6015
Gaithersburg, Maryland       20877

Telephone     202-275-6241

The first five copies of each report    are free. Additional   copies are
$2.00 each.

There is a 25% discount      on orders for 100 or more copies mailed to a
single address.

Orders must be prepaid by cash or by check or money order made
out to the Superintendent of Documents.