Teaching GIS in a Heterogeneous Networked Computer Environment

Geographic information systems are playing a significant part in higher level educational curriculums. Recently at the School of Renewable Natural Resources (SRNR), The University of Arizona, several courses have been developed to exploit the academic and institutional benefits of running a successful GIS educational program. Rapid changes in hardware, software, and operating systems, pose significant challenges in maintaining a quality educational experience for students from various disciplines and backgrounds. While GIS excels at an interdisciplinary level, the job to teach a wide array of students from technologically disparate backgrounds becomes very difficult. This is especially true in a modern heterogeneous computing environment. This paper focuses on the development of an instructional capacity in GIS through the insight derived from a research environment.

Geographic information systems are playing a significant part in university level educational curriculums. Recently at the School of Renewable Natural Resources (SRNR), College of Agriculture (COA), The University of Arizona, several new courses have been developed to exploit the academic and institutional benefits of providing a GIS educational program.

The role of GIS at the university level is different than for community colleges or technical schools. GIS is a tool that provides the student with additional resources, in much the same way as statistical programs. As educators, we need to provide the basic understanding of these tools so that the students will know how the technology will benefit them in their chosen discipline.

As with any emerging technology, implementing GIS courses has been challenging and sometimes nerve racking.

Early attempts at GIS instruction

In 1986, the COA and SRNR developed an instructional computing facility in SRNR, based on Intel 8088 architecture. The lab was used for computer aided instruction in a variety of courses. Applications included word processing, statistics, programming, CAD and GIS. The GIS instructional software for data capture and format conversion, MAPIT, was developed by one of the faculty in the Landscape Architecture program. GIS analysis software was provided by C. Dana Tomlin's Map Analysis Package, and, later, IDRISI. These packages were used to support GIS instructional modules in landscape architecture, as well as an introductory course in GIS. Originally, instruction in GIS was viewed primarily as a component of the Landscape Architecture program and not of much benefit to the general educational function of the College. It would take the development of a computer research laboratory to change the role of GIS in the COA and SRNR.

Experiences with research computing

The Advanced Resource Technology (ART) Group was formed in 1988 within the School of Renewable Natural Resources. ART was created as an interdisciplinary technology research center. Although initially conceived by members of the faculty of SRNR, ART included faculty from various departments including, mathematics, arid lands studies, geography, and psychology. The original intent was to develop a research facility to design and development of computer applications in GIS. With support from the Dean of the College of Agriculture, federal support, and research grants, the lab has progressed to a state-of-the-art computer research facility. ART provides support in the use of advanced digital technologies such as geographic information systems, global positioning systems, high performance computing, and artificial intelligence for applications in a wide range of disciplines. The group provides support in instructional, research, and extension activities within the College of Agriculture.

The activities of the research lab pushed the use of computers and related technology within the School. The need for network access resulted in connecting the building for internet access. The faculty and staff of ART, seeing the advantages to instruction, pushed to have the classrooms and the teaching lab connected.

As the ART facility grew from a PC based to Unix based environment, the mix of PC and workstations required the development of management strategies for heterogeneous computing. The large disk storage capacity in the research lab was seen as a means of handling some of the requirements in the instructional lab. The ART research faculty became the driving force in developing more advanced computer based instruction as well as implementing better GIS courses.

Carry over from research to instruction

As the research lab continued to thrive, resources were diverted to the teaching facility. The instructors wanted to have access to the same types of software that they used in the ART lab. Slowly, changes in the instructional hardware were implemented. Network access was completed and the software was upgraded in the instructional facility.

These changes resulted in better instructional capabilities and at the same time, students were becoming aware of the usefulness of GIS. The student demand for GIS based courses was beginning to exceed the capacity of the School. Students from all over campus, representing such diverse programs as architecture, anthropology, and archeology, joined those in the natural resource areas in a general quest for more GIS courses.

Over a five year period, two GIS courses handling about 60 students per year, became the core of a set of related courses. Additionally, over six courses used GIS for at least part of their curriculum. The demand for a technologically based graduate degree resulted in the creation of an ART option in the Renewable Natural Resources Studies major in the SRNR. Recently, the Interdisciplinary Committee on Remote Sensing, which oversees a University-wide minor in remote sensing, voted to change the minor to Remote Sensing and Spatial Analysis in order to include GIS.

Experience has shown that the human resource requirement for GIS instruction should include backgrounds in geographic analysis and geographic data processing. The same can be said for learning GIS applications. The conceptual foundations for learning GIS applications include an understanding of diverse fields such as cartography, geodesy, database design, and statistics. In addition, general computer literacy can have a major impact on the success of GIS instruction. For example, it became apparent that GIS applications were not an appropriate first computer application for students who have no previous computer experience. The learning curve and subsequent down time for the student who has no computer literacy is not suited to the pace of an introductory GIS scope and sequence. We found that most students cannot recover from such computer literacy deficiencies and keep up with the GIS coursework.

Closely related to instructor issues are those related to instructional database development. In general, most of these issues are related to characteristics of the data set that is to be used in an instructional capacity. Like any GIS database, the exact specification of the instructional data set should be derived from a concise understanding of the functional requirements of the course. In the case of instructional data sets, many of these issues are directly related to functional demonstrations of common GIS concepts (e.g. projection transformations, reclass operations, attribution, etc.) Other issues are related to specific computing environment challenges such as institutional storage, processing speed, and network performance.

The specific functional requirements for GIS instruction will vary from location to location, but generally will affect database characteristics such as geographic extent, feature resolution, categorical resolution, data normalization, thematic depth, projection, and coordinate systems. Categorical resolution refers to type discrimination in related database tables, as opposed to the more familiar spatial resolution, which describes the minimum distance between coordinate pairs in a vector database. Data normalization, here, refers to the atomic nature of logical database domains. Thematic depth is used here to indicate the variety of thematic layers that are available for analysis within a given geographic extent, and, as such, represent a cartographic model.

Needless to say, some requirements for instructional GIS databases are a reflection of the institutional computing environment. Functional requirements, which drive such characteristics as geographic extent and thematic depth toward an educational experience, may be at odds with available disk storage and network capabilities.

Current Implementation Strategies

Human Resources

In response to the increasing demands for GIS instruction at The University of Arizona, ART has developed a strategic approach for providing a quality educational experience. This strategic approach is based on conceptual solutions to the previously describe issues in GIS instruction and guides the programmatic development within ART. We will briefly discuss some of the issues.

The instructional goal of the ART Group is to provide comprehensive quality education in the use of advanced technologies for natural resource management. In order to meet this goal, positions must be created within the organization to deal with a variety of human resource requirements. For example, in computer systems, there would ideally be staff support for systems administration, database management, and network services. Support would be provided within the complex computing environment, with a focus on maintaining the integrity of the local area network function for research and instruction. A parallel function exists in software administration, which has focus areas in version upgrade, licensing, and integration issues.

Given that one of the aspects of GIS instruction is software training, it is recognized that instructional staff time will have to go to the design, implementation, and maintenance of training materials, software, data, and evaluation. Generic texts about GIS do not generally meet the requirements for instruction. They tend to offer overviews of concepts but are not linked to specific examples that are transferable to the instructional database. Instructors need to develop tutorials and other supporting materials based on the specific needs of the course and the databases being used. Revisions to existing coursework, such as measurements classes, require training of existing faculty and/or guest lecturers. Courses that have multiple laboratory sessions with combined lecture may require the presence of student teaching assistants. There is also a growing concern that instructors be certified in various software packages. The complexity of GIS programs requires that the instructor be able to address aspects of the use of such programs beyond the needs of a particular course.

Challenges in student GIS education are being addressed through the development of a sequence of integrated courses with established prerequisites. The integrated coursework will be designed to provide students with progressive, strategic instruction in GIS concept areas. Computer literacy is normalized through prerequisite course requirements that cover operating systems, networks, and common applications such as word processing, spreadsheets, RDBMS, and desktop mapping.

Computer systems support is currently being provided by a combined effort of faculty, staff, and student employees. Key ART faculty within SRNR directly supervise the support of ART's geographic analysis system, instructional computing facility, and network operations. The effort has been greatly facilitated by the creation of two full time staff positions, a system support analyst and a GIS coordinator. These positions are funded by the College of Agriculture. The system support analyst's primary duties include the maintenance of the instructional computing facility, the ART research facility, and general systems support for SRNR. The GIS coordinator's responsibilities, in this area, include GIS software management and support.

Instructional Databases

The provision of quality GIS education requires the development of instructional databases that are tailored to the conceptual content areas of a course and are compatible with existing computing resources. Since its beginning, ART has been involved in the creation of high resolution, multi-thematic GIS databases for research purposes. Such databases can can form part of a GIS instructional data set. However, we have not found any one research database that can solely support any GIS course without modification, update, or supplementation from other data sets. Research data requirements frequently include the need for smaller scale geodata sets for purposes of analysis and/or display of research data.

What is required is on-line access to an integrated, normalized geodata warehouse. The warehouse contains research databases and institutional databases that can be used to synthesize instructional data sets for specific instructional objectives. Such a data warehouse will provide a common location for researchers, instructors, and others from which to retrieve data. Therefore, these data will be of common format, projection, coordinate system, with complete metadata documentation. Furthermore, these data will need to be available to ART researchers and instructors over the network in real time, and available to others by standardized data request procedures.

Currently, instructional database requirements are being dealt with at ART in accordance with the previously mentioned strategies. ART has developed an institutional GIS database which it has been storing and serving data for research and instructional applications. Additionally, ART has become the GIS data service for The University of Arizona. Serving these data needs has lead to the development of the ART GIS Library which contains two types of normalized data. One type is filed into our general reference area, which contains approximately 25 statewide themes at scales of capture as large as 1:24000, with most at scales of 1:100000. Another area in the library contains a collection of project based data sets representing discrete locations around the state, especially in southern Arizona. These sets tend to be at larger scales, and more recent dates of capture than the general reference themes. Many of these sets contain substantial thematic depth and lend themselves to GIS concepts like cartographic modeling.

The ART GIS Library is constantly maintained and updated by data request transactions from other academic units on campus. Each time a request is made, ART acquires thematic information for specific areas, normalizes the format of these data for the end user, and keeps a copy of the data set in the library for later use. This not only reduces data requests from the University to the Arizona Land Resource Information System (ALRIS), but incrementally increases the holdings in our library for instructional purposes.

Such additions are critical, as we have found that each GIS course will require the development of one or more custom databases. It has proven to be unlikely that any one research or general reference data set will be adequate for all aspects of GIS instruction in a given course. Therefore, the design of new course scopes implicitly involves the development of instructional databases that are tailored to the needs of the course. This development is greatly facilitated by the ability to synthesize custom data sets by the integration of normalized research and general reference data in the ART GIS library.

Recently, the ART GIS library has expanded its holdings and functions as part of a cooperative venture in the National Biological Service's National Biological Information Infrastructure (NBII). The NBII will lead to the development of high resolution natural resources databases for National Park Service areas in the state. The initiative will enhance GIS data service capabilities at ART through the development of automated GIS data browsing and downloading functions. Ultimately, this will lead to use of advanced data management architectures using facilities such as Esri's ArcStorm and the spatial database engine (SDE). It should also be stressed that many of the databases used for instruction, have been developed through research funds from cooperating agencies (e.g. National Park Service, National Biological Service, Pima County, U.S.D.A Forest Service, U.S.D.A Agricultural Research Service, and Department of Defense). Without the active research program the availability of databases for education would be limited. Research applications (Pereira and Itami, 1991; Hu and Guertin, 1991) have also led to the development of excellent class exercises.

Hardware and Software

The instructional facility in SRNR has developed into a model high performance teaching lab. The Intel 8088 machines were first replaced by Intel 286 and later 486 machines. The lab is now upgrading to 20 Pentium based machines, running Microsoft Windows NT. There are 10 Calcomp digitizing tablets available for data capture.

The lab has two SUN Microsystems Sparc based machines that act as print and file servers. All the machines are routed through a switched 24 port hub before going out to the main network. Dealing with the various operating systems and getting them to interact has lead to some ideas about what works and what doesn't. In the table below, we compare some of the features of various operating systems that we have used in the teaching environment. This table is meant as a general comment about our experience in trying to make machines with different operating systems interact. Our principle mix at this point in time uses Unix (Solaris 2.5) as the main system, with Windows NT 3.5 as the operating system of choice on the PC platforms.

NOTES:

1. HP makes a device that attaches externally to your printer, directly connecting it into the Ethernet. On the JetDirect EX Plus3 model, the RJ-45 port was consistently losing its connection.
2. You should be able to "capture" an LPT port and assign it to a network port, but this procedure frequently froze the system.
3. Additional software is necessary.
4. The NT system serving the print spool should be at least a Pentium 100 MHz.

Management aspects

The teaching lab is run in much the same way as the ART research lab. The Sparc based servers provide the primary network delivery system for software and disk drive space. The network uses the Unix networking facilities running TCP/IP. This makes linking to other networks easy and reduces complications due to network protocols.

We have developed over the past several years, some guidelines that help to reduce management problems. The following points summarize those guidelines.

i. Invest in the appropriate hardware

GIS courses, and even those that use GIS related materials, require substantially different hardware than courses that use wordprocessing software or spread sheets. The PC used in a GIS instructional facility now rivals the workstation in power and disk space. Additionally, the GIS lab must be fully networked to allow for ease of software installation and access to databases.

ii. Software selection

The use of specific software packages tends to make the student consider one package superior to another. It is important, in the appropriate framework, to expose students to a variety of software packages so that they develop an understanding of the concepts of GIS and not simply a particular vendor's approach to problem solving. However, exposure to specific software products is required to develop adequate skills and a comparative knowledge base.

iii. Accessibility of the computer lab

The specialized nature of the hardware and software requirements for GIS instruction make it unlikely that there will be many accessible facilities. Most universities have a single lab dedicated to GIS instruction. This means that steps must be taken to provide the students with as much free access as resources allow in order to do homework and other functions. The teaching lab is monitored at all times when students are present to provide assistance in case there are problems with the machines or software.

iv. Research and Instruction

The University of Arizona has a unique advantage in that the GIS instruction is driven by faculty in a GIS research environment. This results in research being translated to the classroom very quickly. In addition, the research databases are a wealth of information that can be used in GIS instructional computing. The resources of the research lab also provide access to large data storage arrays and other components of a computer facility not normally available at the instructional level.

v. Licensing agreements

Teaching labs require enough software seats to make it cost effective to seek site licensing agreements. Most software firms will provide some financial break to educational institutions for instruction.

vi. Scheduled replacement

Equipment and software must be maintained at a level that is close to the most current products. Hardware should have a life expectancy of four years, and therefore plans should be in place to assure money at that time. If software is under a site license, the upgrades are usually included in the agreement at no additional charge or for a minor fee.

vii. Partnerships with industry

Private industry is becoming more aware of the importance of education and the role they can play. Our experience has been that major corporations are willing to provide substantial discounts to support instructional activities. The benefit to the institution is access to the very best in hardware and software, that might have been out of reach. Industry partnerships require hard work and commitment, but the payoff is worth it. Some corporations that have displayed a commitment to education are Esri, Intergraph, Motorola, GeoResearch, SUN Microsystems, and Hummingbird. This is not an endorsement of any particular company, but is intended as an example of the diverse support available to educational institutions.

viii. Institutional support

The successful implementation of a GIS curriculum requires a concerted effort involving personnel and organizations. In our case, the support of the School and the College are needed to maintain the lab. Other groups who have tried to implement similar teaching laboratories, have been very surprised at the dollar cost to purchase and maintain hardware and software at the level required for adequate instruction.

The support of the faculty is also necessary. As the faculty become aware of the benefits of GIS to their students and their own research, the teaching lab is seen as a logical extension of the teaching process. The dollar cost mentioned above comes from a finite pool of money. The faculty must support the teaching laboratory if it is to continue to provide the students with the type of educational experience they have come to expect.

Conclusions

The experience we have had in developing and managing instructional GIS at the university level has proven valuable in many ways. The need for cooperation between faculty and administrators to achieve a common goal requires a willingness to compromise. The success that we have achieved in integrating advanced technology into the curriculum has shown payoffs in many ways. The satisfaction of the students is demonstrated by the continued demand for the courses and the addition of new courses. The use of the teaching lab by other faculty to add computer based instruction to their courses has increased.

Teaching GIS as an extension of our research activities has made us more aware of the potential for use of the massive data sets that we are developing. The transfer of our research knowledge to the classroom has kept our courses current and well received by the students.

Literature Cited

Pereira, J.M. and R.M. Itami. 1991. GIS-based habitat modeling using logistic multiple
regression: A study of the Mt. Graham Red Squirrel. Photogrammetric Engineering and Remote
Sensing 57(11): 1475-1485.

Hu, Z. and D.P. Guertin. 1991. The effect of GIS database gridsize on hydrological simulation
results. In: Proceedings of the 1991 meetings of the Arizona-Nevada Academy of Sciences - Hydrology
Section;  Hydrology and Water Resources in Arizona and the Southwest. Vol. 21:59-63.

Author Information

Craig Wissler, GIS Coordinator, E-mail : craig@nexus.srnr.arizona.edu

Michael R. Kunzmann, Ecologist, National Biological Service, Cooperative Park Studies Unit, E-mail: mrsk@npscpsu.srnr.arizona.edu

Dr. George L. Ball, Asst. Research Professor, E-mail: gball@nexus.srnr.arizona.edu

Carolyn A. Audilet, System Support Specialist, E-mail: audilet@nexus.srnr.arizona.edu

Dr. D. Phillip Guertin, Associate Professor, E-mail: phil@nexus.srnr.arizona.edu

Mailing address: Advanced Resource Technology Group, Biosciences East 203, School of Renewable Natural Resources, The University of Arizona, Tucson, Arizona, 85721

Many organizations have contributed directly or indirectly to the development of instruction in GIS at SRNR. The following list is probably incomplete. National Park Service, National Biological Service, National Science Foundation, U.S.D.A. Forest Service, United States Navy, United States Army, Cooperative Park Studies Unit- NPS/USFW/UA, Cabletron Corp., Esri, Intergraph Corp., Motorola, GeoResearch Corp.