By Kenneth J. Dueker, Emeritus Professor of Urban Studies and Planning, Portland State University
Edgar Horwood founded URISA and was its first president.
He provided intellectual leadership toward building an interdisciplinary association and served as a constructive critic in the development of ethical practices in the emerging area of applying information systems technology to urban and regional applications. Those of us fortunate enough to know Ed Horwood can relate to the stories about him. Recounting his encouraging manner, wit, and humor brings back fond memories. Hopefully, this will help others to appreciate his contributions to URISA and the field he initiated and help build.
Edgar Horwood, Professor of Civil Engineering and Urban Planning at the University of Washington, was instrumental in the founding of URISA in 1966. This founding followed several short courses in computer applications in urban planning held in 1962-1965, and several Urban Planning Information Conferences held in 1963-1966. Initially, the conferences targeted alumni of the short courses for follow-up education and discussion of issues in applying computers to urban planning. These conferences led to discussions of the need for a more formal and on-going association of professionals from urban planning, geography, data processing and civil engineering. Existing professional societies and associations were not serving well the emerging interdisciplinary needs for communication.
Horwood’s application of computers to urban planning began in 1960 when he used the first computers at the UW to map land use activity around interstate intersections. This was a Bureau of Public Roads grant to study pressures on zoning at six freeway intersections across the nation. This was followed in 1961 with a Community Renewal Plan (CRP) project in the City of Spokane, and a NSF-sponsored software development project at the University of Washington. The CRP created a need for analysis and mapping of 1960 Census of Population and Housing data, and the creation of a parcel-level database. The software development process responded to these needs. A suite of four programs was programmed in Fortran for the IBM 709 computer. The first two programs were designed for small data sets of 10 to 100 areal units for mapping data at the census tract level. The second two programs were designed for large data sets of 1000 to 10,000 areal units, represented as points for mapping data at the census block level.
The first two programs were called the ARRAY and the CARD MAPPING. The ARRAY program rank ordered values for a census tract variable and drew for each tract a bar proportional to the value of the variable. These arrays were used to identify natural breaks in the data that were used to establish ranges for mapping of census tract data using the CARD MAPPING program. The CARD MAPPING program produced a choropleth map using standard printers. Users specified parameters to select variables from the punch cards and performed calculations to produce arrays and maps.
The second two programs were called the DISTRIBUTION and the TAPE MAPPING. The DISTRIBUTION program served to aid in the identification of mapping ranges for a larger number of observations. The TAPE MAPPING program produced symbol maps to represent the value ranges for census block data.
Horwood was an early requester of digital data from the Bureau of the Census, which led to heavy involvement of Census in URISA and in the development of the their data user access program. But in 1961 the census block data were not yet published, nor did Horwood want to keypunch the data again. So he requested from the U. S. Bureau of the Census the Spokane block data. What he received was data on a magnetic tape in a UNIVAC code that was unreadable on the IBM 709. Consequently, he was forced to invest heavily in time and resources to convert the data for use by the DISTRIBUTION and TAPE MAPPING programs.
Clark Rogers, Arnold Rom, and Bill Clark were key players in the development of the software, data, urban planning application, and supporting educational program. Arnold Rom wrote the four programs, Clark Rogers took the lead on the application and education components, and Bill Clark decoded the tapes from the Bureau of the Census. Clark Rogers and Bill Clark were also principal instructors with Horwood in the offering of a short course in Los Angeles in 1962, several offerings in 1963, Pittsburgh, Berkeley, Evanston, New Haven, and the first conference in Los Angeles. Clark Rogers left for the University of Pittsburgh in 1963 and Ken Dueker joined Horwood’s staff. In 1964, Horwood conducted a month-long summer institute version of the short course that was sponsored by the National Science Foundation. Also in 1964, Clark Rogers organized the next conference on Urban Planning Information Systems and Programs in Pittsburgh. In 1964, Dueker moved to Northwestern University where he and Bill Garrison organized the 1965 conference. Horwood and Dueker also conducted the last of the short courses at the University of Michigan in 1965.
During this period there was a shift in emphasis from software development with a supporting educational program aimed at the planning community, to the provision of analytical services. First, it was recognized that the short courses were insufficient to get people up and running, and several requested follow-up assistance. Performing this function proved difficult within the University and Horwood, Bill Clark, Ken Dueker, Charles Graves, and Hugh Calkins formed a firm called Applied Computer Research Corporation to service out-of-state clients. This thrust precluded further development of the four programs and their porting to new platforms, such as the IBM 360. Others did not pick up and develop the software because Horwood and Bill Clark did not want to lose control by release of the source code. Consequently, the initial software development and training program thrust lost momentum and Horwood moved in the direction of new software developments, such as the Street Address Conversion System led by Bob Dial, and its application at the National Capitol Commission in Ottawa and the Dominion Bureau of Statistics (now Statistics Canada).
While Horwood shifted attention from improving mapping software to address computer applications to support planning and administration, and institutional issues of adapting to new technologies, Howard Fisher, who attended the 1963 short course in Evanston assumed the challenge to develop a better card mapping program. In 1964 Fisher released the first version of SYMAP that extended choropleth mapping to contour mapping and used overprint characters to produce solid map symbols. In 1965 he moved from Northwestern University to Harvard where he established the Harvard Computer Graphics Laboratory. Fisher released the source code and emphasized software R&D. ArcInfo descends from the Harvard R&D program.
The Origins of URISA
Horwood authored a paper that is published in the Proceedings of the Third (1965) Conference on Urban Planning Information Systems and Programs titled, “Association Needs for the Urban Information Systems Field” that called for the creation of what was formed in 1966, the Urban and Regional Information Systems Association. He patterned URISA after the Regional Science Association, an inclusive interest association, not an exclusive professional society overly concerned with credentials.
It is instructive to review the tables of contents of early proceedings to understand the range of issues addressed. Unfortunately, there were no published proceedings for the first conference in 1963, although there was a presentation by Ed Hearle on urban data banks that stemmed from his work at the RAND Corp. that called for the use of parcel-level data for planning and management. The 1964 proceedings contain a number of papers that examine various aspects of information requirements of the then emerging land use models. The 1965 proceedings introduce applications of remote sensing to urban analysis and stress the importance of small area data, and a landmark paper by Nathan Grundstein on Urban Information Systems and Urban Management Decisions. The 1966 conference was where URISA was formed. The 1967 proceedings reflect a high degree of involvement of the Bureau of the Census in terms of digital data access, and the application of small area data in the Census Use Study in New Haven, led by Caby Smith, and staffed by George Farnsworth, Don Cooke, Bill Maxfield, and George Leyland. The 1968 conference is of note for originating the USAC project of integrated urban information systems.
After serving as URISA’a first president Horwood encouraged his students, Clark Rogers, Ken Dueker, Hugh Calkins, and Charlie Barb to play leadership roles in the Association. Actually, he had a broad interpretation of “his students” and many considered him a mentor. He had a major impact on the association, and he was more amused than disturbed by the constant struggle for leadership and power within the organization between the data processing and public administration interests of USAC and the spatial analysis interests of Census and planners that occurred throughout the 1970s.
This struggle for power in the organization was defused somewhat by the creation of special interest groups (SIGs) that allowed for diversity of interests. Charlie Barb was particularly influential in this regard as he was responsible for organizing several SIGs, geocoding, education, and microcomputers.
The Horwood Style
Meanwhile, Horwood was being pulled in various directions, but he relied on students as extenders to enable him to pursue many interests simultaneously. While Charlie Barb and Hugh Calkins were his urban and regional information systems extenders at the UW in the late 1960s and early 1970s, he had other students working in transportation planning software development, namely Bob Dial and Matt Rapp. He was also evaluating rail transit proposals for the World Bank and revising the undergraduate engineering curriculum. He entrusted his students with responsibility and enjoyed watching them grow in the process.
Charlie Barb writes:
“Ed’s influence upon URISA and the field (was) his personality, intellect and off-center insight. He was so big that there was a recognizable “Horwood School” that we and many others were proud members of -- how many people have such an impact, influence or point of view? There were his memorable baffoonish antics, like hiding under the head table at a URISA conference while he was being introduced -- he really lacked conventional pride and a preoccupation with self-importance.
While preparing a eulogy for Ed at a memorial held at the University, I asked other students who had worked with him what set him apart. Matt Rapp’s comment was the most insightful and memorable: his compassion -- which, in its unique paternalistic fashion, was what set him apart and endeared him to his students. He was a gold standard “professor” and I think that we knew it. (He was a) creative intellect. A characteristic of it was his simultaneous grasp of an issue or topic from multiple perspectives, including the reverse. He was a great challenge to conventional wisdom and politically correctness. During his years, there really wasn’t anyone in URISA that commanded his respect and at the same time, affection. He needs to be described to the URISA membership from this human perspective because there are so many in URISA’s past and present that still miss him, a person who, uniquely, is missed -- these many years after his death. I hope you can some way include these dimensions into the meaning of Ed Horwood to URISA.” (Barb, Charles, e-mail to Ken Dueker, December 27, 2001)
Many remember Horwood's Short Laws of Data Processing and Information Systems:
As suggested by Horwood’s Laws and the quote from Charlie Barb, he loved to poke fun at people who took themselves too seriously. Yet, he undertook to create a serious interest association of professionals, but not a professional society. He defined a professional society as one who was more interested in keeping people out, by certification. I wonder what he would say about the certification of GIS professionals?
On the other hand, he would be quite enthusiastic about the GIS Code of Ethics. He would see this approach as more central to the ethical issues and choices, and at the same time seek common ground between certification and the code of ethics.
He would engage in such a debate with insights and humor that would cut to the heart of the issue. Although he was concerned with “charlatans” in the field, he preferred to unmask them with challenges and ridicule. He appreciated good work and was quick to praise it. Similarly, he was quick to challenge sloppy work and thought; sometimes by interruptions with humorous or embarrassing comments.
Horwood would be a strong proponent of metadata and the NSDI spatial data infrastructure concept. However, he would be critical of associated hype and claims. He would be suspicious of the strong federalization of spatial data infrastructure. He would be a proponent of fostering local governments to develop enterprise-wide data models.
In this spirit I offer some extensions to Horwood’s data laws to address current issues:
Dueker’s Extensions and Updates to Horwood’s Data Laws
- Meetings about data sharing are more time consuming and costly than the time and cost of data duplication.
- It takes longer to build the infrastructure database than it does to build the infrastructure.
- Surveyors can’t do anything with inaccurate data, and planners can do anything with inaccurate data.
- GIS analysts can overlay inaccurate data with great precision, which makes good data out of bad data.
- Stovepipe systems are used to cook data.
- Data sharing is politically correct, but not practiced.
Those of us fortunate enough to know Ed Horwood can relate to the stories about him. Recounting his encouraging manner, wit, and humor brings back fond memories. Hopefully, this will help others to appreciate his contributions to URISA and the field he initiated and help build.