textfiles/law/p_view.law

April 1991                                                        

                                                                  
                         POINT OF VIEW                            
             A MANAGER'S GUIDE TO COMPUTER PROJECTS                            

                               By
 
                        Charles Brennan                      
                           Inspector 
                 Information Systems Division 
          Philadelphia, Pennsylvania, Police Department             

                                                                  
     Increasingly, police managers find themselves placed in 
charge of computer projects within their departments.  Many feel 
ill-equipped for the task because, although they know the 
operational side of their agencies very well, the technical 
aspects remain, in large part, a mystery.  It is easy for even 
the most attentive manager to become lost at meetings,
understanding little of the technical jargon and having to make
decisions without a full understanding of all the facts.
However, this need not be the case.  Managers can take a number
of steps to help ensure that technical projects will be
completed successfully.

PROJECT MANAGEMENT                                                

     A large-scale technical project requires a project manager.
This person should be of sufficient rank or standing in the
department to make almost any decision independently once the
job begins.  A full-time manager is best, and many
hardware/software vendors suggest that the department assign an
individual solely to this task, since the project manager is the
primary contact between the vendor and the department.

TECHNICAL KNOWLEDGE

     Those assigned to a computer project who have no technical
expertise at all must develop at least a working knowledge of
computers.  A recommended source is introductory courses at a
local university.  In addition, there are many books that
explain both computer terminology and the operational aspects of
the computer.  A thorough understanding of the terms is
especially important, since project managers must be able to
comprehend what is discussed at meetings that they will be
required to attend.  But no matter what the technical
background, it is important to make sure that meetings with
vendors operate at the project manager's level of understanding.
Vendors want to retain customers, and therefore, will take the
time to explain the technical aspects in laymen's terms.  If any
fail to do so, the best course of action may be to consider
another retailer.  However, such issues should be dealt with
before issuing a contract for products or services.

     For large and complex projects, however, it is advisable to
have some independent technical assistance.  If the department
has an in-house technical staff, they may provide all the help
needed.  If not, the city or county may have technical resources
available.  In any case, the technical team assembled should
work closely with vendors providing equipment and software.

PROJECT GOALS                                                     

     Every project has goals that must be satisfied, and computer 
projects are no different.  The project manager must understand
each one and how they affect separate entities within the
department.  For example, in large departments, the installation
of a computerized records management system must satisfy the
specific needs of many departmental units.  It is important to
realize from the outset that the system eventually put into
place may not satisfy everyone's expectations.  But, one of the
manager's most important duties is to meld all of these
seemingly competing needs into what is both practical and
possible for the entire department.

ADMINISTRATIVE SUPPORT

     Department administrators must understand and support the
project.  The head of the agency will certainly be aware of the
implementation of a large-scale computerization project, but may
not fully understand the impact the project will have on the
department.  For example, the installation of a computer system
may require changes in departmental procedures, personnel
allocations, and other fundamental aspects of the department's
operation.  The chief administrator should be kept aware of the
project's progress, as well as what departmental changes will be
necessary, through regularly scheduled status meetings.

TIME TO PLAN

     Proper planning is probably the single most important
factor to computerize a department successfully.  For every hour
spent in good planning, 10 hours of aggravation can be avoided.
The project manager should take time to enlist the assistance of
employees and designate tasks.  One way would be to form a
committee to guide the project.  Another consideration is to
assign tasks, responsibilities, and timetables so that everyone
knows what jobs must be done and who must do them.

     Potential repercussions for the agency should be
anticipated.  For example, the introduction of personal
computers (PC's) in an agency goes beyond just buying the
machines.  Such a purchase raises questions:  Who will fix the
equipment?  What about training?  Are there certain procedures
governing their use or the information they contain?  Other
issues regarding specialized programs and additional software
must also be addressed eventually.

     Every computer system needs support that requires resources
and personnel.  And, while a computer system may reduce the
number of people needed for a certain task in one area, it may
increase the personnel required in another.

RESEARCH

     In any computerization project, there will invariably be
problems encountered and problems to be solved that have the
potential to be overwhelming.  New software packages
particularly are subject to problems.  Testing in the "lab"
cannot adequately duplicate real life conditions.  Therefore, if
an agency is the first to install a package, it may be faced
with complex problems to which there are no known solutions.
Essentially, the department will be tasked with solving the
problems for all departments purchasing the package thereafter.
This could prove to be not only inconvenient but also very
expensive.

     There are two conditions, however, in which this general
rule may be disregarded:  1) If a vendor offers a substantial
discount, or 2) if the software is so unique and innovative that
it cannot be purchased or tested elsewhere.

USER INPUT

     One of the biggest mistakes that could be committed in any
technical project is not to involve the users in all phases of
development.  No one understands the job better than those who
have been doing it for years.  Shortcuts that have been
developed over time to "work around" problems could be missed if
no one is consulted on "how it is done."  And, making changes
later to accommodate these procedures could be expensive.  Only
by involving those who will use the program can project teams be
certain to develop a system designed for the job.

THE RIGHT VENDOR

     If planning is the most important facet of a technical
project, then choosing the right vendor is next.  Although there
are no rules for vendor selection, there are certain guidelines
that should be followed.

     *  Contact Other Departments

        Every potential vendor should supply a list of clients
	who have installed similar systems.  The project manager
	should take the time to contact a random sample of these
	clients to ask questions regarding the workings of the
	programs, problems encountered, and advantages and
	disadvantages of the system.  The project manager should
	prepare a list of questions so that all important points
	are covered.  Most departments are willing to share
	information about the reliability and performance of
	vendors.

        Some vendors may claim that they are "business partners"
	with a larger computer firm; however, this does not mean
	that the larger company guarantees the vendor's products
	or software.

     *  Make on-site visits
                                               
        The best way to see if vendors can do what they promise is
	to make a site visit.  Before selecting a product or
	service, it is important to see it in operation.  If at
	all possible, the police department using the product or
	service should be comparable in size to the one
	considering the purchase.  Again, the project manager
	should plan for any site visits by formulating questions
	and determining what functions are important.

     *  Judge vendors by the same criteria

        It is important to judge every software package or
	product by the same standards.  This can be done by
	creating a matrix with vendors listed down one side of
	the page and the different criteria listed across the
	top.  By placing an "X" in the column where the vendors
	meet the criteria, the project manager will have a
	simple and easy method to evaluate vendors and to
	determine which ones meet the standards necessary for
	the project.

PROBLEM LOG

     All problems encountered in completing a major technical
project should be relayed to one individual.  This individual
should record the problem, who reported it, and how, or if, it
was resolved.  Among other things, a problem log ensures that
all issues are communicated to the vendor centrally and in the
same format each time.

SAFETY NET

     In a complex project, it is very important to construct a
type of safety net that would anticipate upgrades and factors
overlooked during initial program development.  This is
especially true if the project involves the purchase of
software.  A good safety net for this type of project is to
include a provision in the contract that requires the vendor to
provide a certain number of programming hours to "enhance" the
purchased software.  In many cases, the enhancements are changes
required due to circumstances not anticipated when the software
specifications were given to the vendor.  It is imperative to
remember that certain manual procedures may not translate easily
or cleanly to an automated format.  The larger the project, the
more likely things will be missed.  Unless provisions are made
ahead of time, changes to the original specifications may
require additional resource outlays.

IMPLEMENTATION                                                    

     Once a new system is implemented, there is the impulsive
tendency [to] "get it up" and have everyone using it
immediately.  This tactic usually only confuses the users and
breeds frustration.  If the system has many different
components, it is a good idea to introduce them gradually, to
ensure that all users are operating at the same level of
understanding before moving forward.

     If possible, it is also advisable to test the system in a
small segment of the department before releasing it for general
use.  This will allow time to gauge the reactions of a small
sample group in a controlled area.  It is better to find
problems here and correct them before everyone begins using the
system.

CONCLUSION                                                        

     Technical projects require planning and forward thinking. 
But, even the most complex projects can be successfully
completed if certain guidelines are followed. While some
technical background, or at least familiarity with computer
terminology, is important, a comprehensive, well-defined
approach is invaluable to complete a technical project
successfully.

---------------
     "Point of View" is a forum for law enforcement
professionals to suggest recommendations to improve police work.
Submissions for this feature should be no more than 750 words,
typed, double-spaced, and forwarded to Editor, FBI Law
Enforcement Bulletin, Room 7262, 10th & Pennsylvania Ave., NW,
Washington, DC 20535.