- Institutional Relationships
You must continue to establish or strengthen institutional relationships that will support the GIS effort. This was begun with the establishment of the GIS committee and the initial building of political support. It may now take the form of letters of understanding between departments and organizations making commitments of organizational support, staff time, data, and/or funding.
Also build relationships with those organizations and departments within and outside of your unit of government that stand to benefit from the GIS and can contribute to its success. For example, Dakota County was able to defray much of the cost of its excellent parcel-base by working together with the local electric utility. Both organizations benefited from the leveraging of each other's strengths. In less densely populated areas, several jurisdictions can join efforts through regional organizations to achieve the critical mass of needs and resources to make a regional GIS development successful.
GIS will change your organization, just as every new way of doing things has changed it in the past. Change can be uncomfortable for many people, particularly those who excelled at the old way of doing things and rose in the organization because of it. A new technology can make practical sense and offer many benefits, yet still not be accepted. It may make too many staff members or managers uncomfortable, changing their jobs and their relationships with each other. Just as for elected officials, build acceptance among of managers and staff members by offering information and involvement and by finding ways GIS can aid their favorite projects.
The world around us will cause your organization to change, regardless of GIS. The trick is to get on the wagon of change and help steer it, rather than be run over by it.
Establishing and solidifying institutional relationships is one of the most critical and difficult steps in the entire process. Strong support from a few organizations or departments can be more important to GIS implementation than lukewarm support from many. The key is to get one or more departments actively involved and the others at least not actively in opposition. You may find it useful to involve a consultant who specializes in organizational dynamics and team building.
- Service Delivery
Thus far you have consulted potential beneficiaries of the GIS, determined their needs, investigated GIS resources necessary to meet these needs, refined your overall goals, and set objectives. Now it is time to determine the kind, amount, and distribution of GIS services that will be provided and who will be responsible for providing them. In consultation with the Users Committee, the Policy Team should consider questions of service allocation. For instance, what projects or departments will have priority when GIS resources are not adequate to meet all needs at the same time? What will be the consequences of not satisfying requests for service?
The answers to these questions are key to maintaining and building support for the GIS throughout the organization as the implementation progresses. Frustrating early supporters and failing to serve those who want to become supporters will be more damaging to the implementation program than will failing to interest these individuals and departments to begin with.
If your implementation is organized to benefit only a subgroup of early supporters, and/or constrained from providing services to those who are likely to become interested, it will be in trouble. Avoiding these problems may require special service arrangements which provide staff or budgetary support from the various departments expecting services. It may also involve some form of expectation control. Unfortunately, GIS is "in fashion" these days. This too often leads to overselling and inflated expectations.
- Software & Hardware Overview
Hardware and software acquisition are too often the primary focus of GIS planning. GIS needs and objectives should be the primary focus and should guide hardware and software acquisition. In general, computer hardware gets cheaper for a given amount of processing power as time goes on. This money saving trend is offset to a certain degree by the increasing processing demands of the software, as more features and capabilities are added.
Change is rapid. There are drawbacks to buying everything at the beginning. By the time you need that additional computer capacity it may be selling for half of what you paid for it. Additional software functions may be superseded by something better by the time you finally get to use them.
However, waiting for the next version of the hardware or software can also be unsatisfactory. There will always be a better version tomorrow, but you need to get work done today. The key to investment in such a dynamic technology is to buy what you need, when you need it. Pay attention to upgradability and the potential for learning and datasets for the current system being useful in likely new ones.
Ask these questions to help avoid slipping into a backwater, from which escape may be expensive:
- Is the computer upgradable for a reasonable cost?
- Can it be expanded to support additional Central Processing Units (CPUs) and more peripherals, such as disk and tape drives?
- Is it well supported by the manufacturer and dealer?
- Does it meet organizational standards for hardware and software? That is, will it work with your other systems, or will it be an island of technology?
- Does it use standard technology that is well understood in the technical community?
- Is the GIS software well supported technically?
- Is the maker a stable firm that is likely to be around for a while?
- Are there other users in the area who can provide advice?
- How hard is it to hire technicians who know this software?
- Is the data model (or format) widely understood and easy to translate into and out of other major formats?
- Are the data and learning investments of a small installation preserved when you want to move up to more powerful versions of the software, or will you have to reconvert your data and start from scratch to learn new software?
By asking these questions, you can stay more or less in the middle of the stream of technological change and avoid the sloughs and backwaters.
Computer systems can be complex and are mostly beyond the scope of this guide. Here are several basic issues that you should consider and discuss with your computer staff and hardware & software vendors.
- Software & Hardware
- Hard Drive Capacity
Get enough hard drive capacity to accommodate your software, your data, new data you will obtain or generate, and the temporary files needed during processing. Needs will vary greatly, but at minimum in the 1996 market, do not buy a PC for GIS processing without 2 GB of hard drive capacity (half of this for a machine running a GIS viewing software). Figure twice this amount for a UNIX workstation. Keep in mind that you may need temporary file space on your hard drive that is as much as four times the size of any GIS data set you are processing.
- Backup and Data Exchange
You will also need the ability to archive data and efficiently exchange large quantities of data with others. This probably means a tape drive of some kind, although high capacity removable disks and writable CD-ROM technology are falling in price and can give better performance. Pay attention to what your likely cooperators and data sharing partners have. If you will exchange data with them, be sure you have the right hardware and software to do it.
- Processing Capacity
Get a fast enough Central Processing Unit (CPU) and enough Random Access Memory (RAM) to give you the processing horsepower you will need. For GIS processing do not buy a PC with anything less than a 133 mhz Pentium with 32 mb of RAM. A 486/66 with 16 mb RAM would be the minimum CPU for GIS viewing software. If you are buying a UNIX workstation, ask the vendor to demonstrate the computer running the GIS software and a data set of realistic size to judge how much capacity you will need.
- Video Display
The video display subsystem should include at least a 17" color monitor and a good video card with 2 mb of RAM or VRAM. The quality of displays varies, and your technician will spend his life looking at it. Preview the display before you buy, if possible.
- Plotters and Printers
Plotting and hard copy map creation will require some sort of color printer or plotter. Quality has gone up and prices down on these items, so you can get some good equipment much cheaper than before. These days, ink jet technology rules. These plotters are really very large ink jet printers. They are good for both line work and color fill. However, they go through a lot of ink and will require a large amount of RAM to do large color fill plots. If you are doing color fill maps up to 36" x 48" you will likely need more than 60 mb of RAM in the plotter. For 8 1/2 X 11 output, ink jet and laser jet printers perform well.
- PC vs. UNIX Workstation
This choice will depend upon the requirements of the GIS software you have chosen. This software choice will depend upon the software functions needed, the size of GIS datasets, the size of the land areas they cover, and the nature and extent of user access to the GIS. In general, a good UNIX workstation will cost you at least twice as much as a top of the line PC. Moreover, UNIX systems are more difficult (i.e. expensive) to maintain than PCs. Lower end UNIX workstations are not necessarily faster than high end PCs doing the same work. UNIX workstations tend to be better made than the highly priced competitive PCS but, in any case, most PCs are obsolete before they wear out.
UNIX is designed for multitasking and distributed computing. It is a 32 bit (or increasingly, 64 bit) operating system as opposed to the 16 bit DOS/Windows PC. A 32 bit operating system can handle twice as wide a piece of data at one time as a 16 bit can -- thus more speed. But with the advent of 32 bit PC multitasking operating systems designed for distributed computing, such as Windows NT, these distinctions are beginning to blur. UNIX workstations still have some advantages, but the PC world is catching up.
An investment in UNIX based GIS software running on UNIX workstations will probably provide all the power you will need, but at a high cost in money and complexity. An investment in a PC based GIS will be less costly and complex, may still do most or all of what is needed, and can be a good step along the way to the more sophisticated UNIX based GIS later, if justified.
Our general advice is to stick with the PC if you can. The major GIS software vendors have or will soon have PC versions (via Windows NT) of their most powerful software packages.
- Systems Topology
For purposes of discussion we designate computer hardware and software in a GIS as either "GIS processing" or "GIS viewing" hardware/software. "GIS processing" hardware/software refers to the system on which the digital maps are made and maintained and on which higher level topological analysis is done. "GIS viewing" hardware/software refers to the systems on which more casual users query the data and make simple thematic maps.
Your GIS processing computer systems can be configured for GIS in one of three ways:
- Stand-alone PC or UNIX workstations are the simplest and cheapest configuration for small organizations or early implementation phases. Many need never move beyond this. Such configurations maintain critical data files and support processing locally. Several stand-alones can be linked together for data transfer purposes with a peer to peer network. Such configurations are relatively cheap, simple options to start out with. When needs increase beyond the range of this configuration, the investment in the stand-alones is preserved by using them as components of more complex configurations.
- In file server/workstation installations, one central computer contains the majority of the data and is accessed over a network by a number of computers which locally handle the GIS processing of data. The advantages of this configuration for larger GIS operations is the single storage of GIS data sets that are shared by several technicians. This cuts down hardware requirements, since the same GIS data sets are not being stored in many locations, and it helps ensure that everyone is working with the same version of the data. This configuration is often the first step up from a stand alone system as the GIS technical shop gets larger.
- Very large GIS technical shops may be able to save money by moving to a file server/computer server/X terminal configuration. In this case, most operators use relatively inexpensive PCs to access data on the file server and also perform the processing on a remote computer server. The computer on the technician's desk does little more than provide a way to communicate remotely with the central computers. This configuration is very close to the old mainframe computer & terminal model and shares its shortcomings. When a central computer is down, all computer work stops.
Most GIS users should have GIS viewing software on their Personal Computers and access to the GIS processing computer and other computers with pertinent data via a computer network. The GIS processing computer or file server will be the repository for digital maps and some tabular databases, with other tabular databases residing on computers in the departments that maintain them. GIS is most successful when direct involvement with it spreads beyond the GIS professionals to include those who are specialists in doing the normal work of the departments. NAIS respondents are using their GISs extensively for data query. This kind of use lends itself well to wide distribution among departmental staff via GIS viewing software and computer networks.
Mn/DOT has prepared a report called "GIS Platform Guidelines". This short and very practical report is based upon extensive Mn/DOT experience and practice. It can be helpful for others in government who are planning computer systems for GIS. Contact John Schwartz or Vic Conocchioli, at Mn/DOT in St. Paul to get a copy.
- Software & Hardware Support and Upgrades
Many organizations purchase maintenance contracts or extended warranties for their computer hardware. In many cases these contracts can be more costly than paying for repairs as they are needed, but it can be easier to budget for a service contract than to guess how much your hardware repair costs will be in a given year. More complex computer systems, particularly if they include UNIX workstations, are depended upon by many people and are difficult for users to repair. Though expensive, having a trained technician on call to get the system back "up and running" can be important.
Most organizations purchase software maintenance and upgrade contracts. The software does not wear out, but technical problems and questions do arise, and it is very useful to be able to make a call for help. In addition, many of these contracts include upgrades to new versions of the software when they become available. As in the case of hardware, it is useful to be able to budget for a maintenance contract rather than try to guess in advance when new software versions will ship and how much they will cost.
- Revise Cost Estimates
When the GIS design has progressed to this point, you know how to specify the GIS software you need and the hardware it will run on. At this point you should consult GIS vendors, showing them your plans and asking them to provide detailed specifications and prices for the software and hardware.
Your plan should also address sources of funds for the GIS effort. Most GIS efforts in Minnesota local government are funded through regular departmental budgets. This can cause difficulties handling start up costs such as hardware and software, special training and hiring, and, particularly, base map creation. Be creative. Look for ways to use special funding sources. Explore cooperative possibilities with other units of government (nearby cities and counties), private sector organizations (electric utility, cable TV, forest products companies, etc.), and regional governmental organizations (Metropolitan Council, etc.)