Want to build A GIS? The proof is in the prototype
Meanwhile, provided that funding is immediately available, rework to the prototype (or even
complete replacement) can be undertaken, and iterative deliveries scheduled. All of the data
model changes, higher priority application enhancements, and even many of the lower priority
enhancements that were deferred earlier can be introduced into version 2.0.
Pitfalls to Avoid
The biggest pitfall to avoid and the one most often encountered is not controlling scope. The
best way to avoid this, as I have repeated throughout this paper is discipline and reassurance that
the system does not have to be perfected the first time and that there will be time later for
improvements.
Another pitfall, once the working prototype is developed, is not treating the prototype as such. It
may be very well worth performing a full overhaul on the system to reduce the longer term
maintenance costs of the system. If the data model is flawed, this is the time to correct the
problem and rewrite applications, as necessary. The beauty of prototyping, designing on the fly,
and object orientation is that you never really have to throw away everything.
Finally, it is important to use small development teams. The familiar adage that you cannot
speed up delivery of a baby to one month by assigning nine women is painfully true.
Furthermore, to use another clich4, too many cooks spoil the soup. Small teams are much more
manageable and effective than large ones. An experienced software development team should
require no more than twelve people.
Examples
Following are three real world examples that show where rapid prototyping has been utilized to
realize quick, fast return, high visibility results early in the GIS project life cycle.
Case 1. Lame Electric Utility
At a large northeastern U.S. electric utility, many of the techniques described in this paper were
employed to great success. Upon selection of a GIS platform, the project team immediately
embarked on a rapid prototyping effort that involved creation of a throw away data model,
utilized existing digital data sources, enabled the support team to become experienced quickly,
heavily involved the SME’S, and made several incremental deliveries in the span of just a few
months. The system was installed at several remote planning sites and, though it was only
simple application, paved the way toward user and organizational acceptance very early.
Simultaneous to this effort, much of the core functionality was developed and much of the
feedback from the prototype system was incorporated into the functional design and core
development. Likewise, much was learned by the development and support teams during the
prototyping phase and many of the software modules developed for the prototype were able to be
reused in the core system. The end result was one of the most robust, working AM/FM systems
implemented to date. The championing and marketing of the system and its concepts throughout
the organization have ensured that the AM/FM system will not only be used to automate but that
it will also be able to introduce real, cost-saving efficiencies to the organization as a whole.
Case 2. Medium-sized Gas Utilitv
At a medium sized gas utility in Canada, in the span of only a few months, the company was able
to implement a working map maintenance application for their drafting department. The system
supported the complete life cycle of gas distribution facilities Erom the time of a “main extension
request” from marketing through to the point where facilities are abandoned. To reduce the up
front time and cost of data conversion, their distribution maps were scanned and raster images
utilized as a backdrop. The success of this project has been due, in large part, to the fact that this
application was rolled out as the first of several small self-contained phases, rather than as a
whole. It was also made more successful due to heavy involvement by the users of the system,
throughout development. Application development was further enabled by use of a GIS
platform with a filly fictional object-oriented language. The project came in under budget and
was able to further justifi itself on the basis of this success and the degree to which the
technology could be used to enable additional functions that were outside of the
engineering/drafting group.
Case 3. Telecommunications Commuw
A northeastern U.S. telecommunications company was able to develop a working prototype of a
broadband engineering design application in less than six months. This was made possible
through the use of very intense prototyping and “designing on the fly” that allowed for maximum
flexibility during the development phase. This resulted in a surprisingly robust initial application
that might have, under other circumstances taken ten times as long.
Conclusion
It is extremely difficult in today’s fiscal climate to justify large expenditures on GIS technology
on the basis of cost-benefit studies alone. With the advent of object-oriented GIS technology that
enables rapid prototyping, the existence of many existing digital data sources, and scanning
technology that allows paper maps to be quickly brought into the system, it is at last feasible to
build working GIS applications from scratch in a matter of weeks where it once took years.
Prototyping and rapid implementation of a “quick hit” high visibility application very early in
the project’s life cycle is the key to both early and lasting success for a GIS project. The
corporate buy-in and extension of the system to other departments in the organization that results
will ensure that the necessary support for the system will be kept up for many years in the fiture.
Most importantly to the visionaries among us, there is the satisfaction of knowing that we were
able to make a difference that actually saved the organization and its customers money and
helped to make the operation more efficient and therefore more competitive in the marketplace.
References
Phillips, Larry, “Eliminating the Multi-Year Nature of AM/FM Projects”, AM/FM Proceedings
1995.
