This web site documents my recollections and experiences during my 50+ years in the GIS and geospatial technology industry. These historical notes are not about the major players and events but more about the many significant but unrecorded and unheralded developments in the evolution of our industry that makes geospatial technology what it is today. I would encourage others to provide me content!
The following narrative and pictures document the path of my career as a geospatial/GIS technology professional.
Below is a picture of the first terminal I ever used. I worked on interactive programs for mapping meteorological observations in Vermont. Thanks to Dr. Vince Malmstrom for getting me interested in computers. This was during my junior and senior years (1969-1971). The terminals were located in small rooms - perhaps old broom closets. About 5 by 6 feet. Believe it or not, Dr. Malmstrom wrote an online Geography concepts tutorial. Perhaps the first Geo e-learning capability ever! I have a copy. Very forward thinking.
At 300 baud, it took a while to generate a line printer map with over-strikes! This terminal was connected to the Dartmouth GE 635 timeshare system.
When I went to graduate school (SUNY Buffalo) I had to use computer cards in my first computer classes! A big step backwards.
That changed in 1974 when Prof. Duane Marble started a GIS lab. His project was able to purchase a Data General NOVA 840 minicomputer and a Tektronix 611 storage tube. I programmed (in assembly) the graphics driver for the CRT. Great fun watching green squiggles become maps. The images below are aa DG NOVA 840 and a Tektronix 611 graphics tube. Dr. Hugh Calkins joined the SUNY Buffalo geography staff at the same time as Duane.
What was pretty cool about this 611 installation was that the Dept of Electrical Engineering at the University of Saskatechewan designed and built a custom hardware graphics board. Thanks to Dr. Ray Boyle's team at the University for that wonderful bit of creativity , we were able to achieve 300K baud graphics rendering speeds. (Read the Early Innovator: Ray Boyle page.) This was a considerable improvement over the old teletype I used at Middlebury and the punch cards I used in my first year of graduate schoo!. What was not so cool was loading the operating system on the NOVA 840 - we had to use paper tape. However, after a year or so, the lab purchased a mag tape drive. Great improvement.
During my last two years in graduate school, I was in charge of the departmental software library. I was responsible for converting software from other universities and research organizations so that the software could be used for class work at SUNY Buffalo. Some of the cooler applications I converted to run on the DG System (as well as the University CDC system) was Dr. Tobler's Cartogram software and a suite of software for doing multi-statistic based classification for geospatial features. This included software for spatial auto-correlation which I used in my Master's paper research.
I was also responsible for developing a set of software libraries that could be used by other GIS programmers in the lab. So, I designed a developed a series of character manipulation, math, and graphics routines as individual Fortran libraries: Object philosophy before I ever heard of the concept. Just seemed common sense. I used many of these libraries in the development of the first version of MOSS (see below).
For the summer of 1974, Prof. Marble arranged for me to spend 8 weeks in Saskatoon Saskatchewan working with Dr. Ray Boyle and the Automated Cartography Lab in the Electrical Engineering Department at the University of Saskatoon. This was my first exposure to DEC equipment: A PDP 8e with 32 Kbytes of memory using, if you can believe that, and 12 bit words! They were doing amazing applications on this machine: Automatic line following, real time digitization, cartographic composition, and so on. Truly impressive and leading edge. As I mentioned before, being in the EE department they would also design and build custom graphics cards to enable vastly improved performance of their software applications. Their lead developer previously worked at Boeing as a lead design engineer.
PDP 8 face plate
My role in the group was three fold:
1. Learn PDP assembly language
2. Learn how to use and exploit the PDP 8 Floating Point Processor.
3. Work on a solution for contouring a set of random 3d sample points.
With some very long days and weekends, I was able to accomplish all three summer objectives. I wrote reports on these activities so I could get course credit. The really interesting aspect of the work is the solution I developed for contouring aq random sample of 3d points: triangles. Seemed the only logical approach. Take the set of points and define a set of triangles and then contour based on the triangles. Unfortunately, I had not heard of Delaunay triangles. Chris Gold used Delaunay triangulation a few years later to define a much more robust approach for generating triangular mesh networks in GIS and other tool software. Anyway, my approach worked. I also coded an interactive edit program for correcting the results of the contouring process.
1974 was also the summer in which I saw my first rugby match! Great sport.
In early December 1974, I attended AutoCarto 1 in Reston Virginia. My registration fee was waived as I and other grad students "volunteered" to help with registration, audio-visual setup and so on. I traveled to the conference with Duane Marble and Hugh Calkins. I distinctly remember driving from the hotel to the Reston Town Center (the Sheridan Hotel) to the Lake Anne center with Duane, Waldo Tobler, and a couple of other early GIS notables - and getting totally lost. Not one of these GIS pioneers thought to bring a paper map! So much for geographers and map reading. At that conference, I also met Jack Dangermond, Bill Mitchell (USGS), and Robert Aangeenbrug (Bureau of the Census) for the first time.
In 1975, Dr. Perry Hanson, my PhD Advisor, convinced me to submit a grant application to the Law Enforcement Assistance Administration (LEAA). To my surprise, I was awarded the grant for various expenses including travel. A step back. In 1973, Dr. Hanson developed a working relationship with the Buffalo police Department. This relationship allowed the Geography Department to have access to years worth of offense and arrest records (names redacted). We received this information on mag tapes! We (Dr, Hanson and a few grad students like me!) had to load all the data onto the mainframe and then process/curate the data so that the data could be used for spatial analysis and mapping. This processing included running US Census software to convert addresses to census tract, block group, and block designations. This geocoding information was then added to the offense and arrest records. A number of Masters and undergraduate students used this processed and curated data in research and masters papers. My focus was on auto the City of Buffalo with a focus on predictive modelling. Modelling and mapping were the two primary application areas that the Police Department was interested in. One of the Masters students at that time was a Sergeant with the Department. His knowledge of police work and data collection practices really helped the rest of us!
I finished my PhD course work in 1977 and moved with family to Colorado. I took a job as a GIS software engineer on a GIS development project (see Early history of GIS in Ft. Collins). My job was to design and implement a new public domain GIS. The majority of the development work was done using a Tektronix 4010 graphics tube and keyboard operating at 9600 baud connected to a Control Data Corporation (CDC) 6400 mainframe running the Kronos time share operating system. This was later upgraded to NOS in the late 1970s. Had to love those thumb wheel cross-hair controllers on the 4010!
I finished beta development of the Map Overlay and Statistical System (MOSS) in 1978. We did initial training and testing through the rest of '78 and '79. During this time, I also had the opportunity to meet and work with Harry Niedzwiadek and George Hoffman, both of Autometric. While I was working MOSS, Harry, George, and Cliff Greve were working on an interactive digitization application that allowed users to digitize from either planimetric maps or from stereo imagery. The software application was called WAMS (Wetlands Analytical Mapping System). This work was being done under contract to the Wetlands group of the US Fish and Wildlife. More on this later, but some interesting technical bits: WAMS supported on-the-fly topology construction, incredibly efficient topology validation and polygon formation, and real time graphics feedback as the user digitized and edited data.
A slight digression in the chronology. In 1977, I was fortunate enough to be invited to write a paper and attend the Advanced Symposium on Topological Data Structures. This Symposium was organized by Harvard University and sponsored by several Federal Government organizations. The actual conference was held in Dedham, MA in October of 1977. My paper was on designing and writing portable software. An interesting aside was a comment I made in the paper about the need for standards in the GIS community. An harbinger of things to come. Hardcopy versions of the proceedings of this symposium, published in 1978, are extremely difficult to come by. This is unfortunate as this symposium represented a milestone in the developmental history of GIS. Abstracts of the collected papers can be viewed here: http://training.esri.com/campus/library/bibliography/Browse.cfm?referencetype=9&LibSection=Conference%20Proceedings&BrowseCategory=Harvard%20Papers%20on%20Geographic%20Information%20Systems&Year=1978. I have a copy that should be scanned and made available on line for free.
In late 1979, the Western Energy and Land Use team completed a new computer facility with raised floor, a Data General Nova C 330, with 128K Bytes of memory and 192 M Bytes of disc (the drive was the size of a washing machine). They also purchased and installed a Tektronix 4014 (still with thumb wheels). This is the terminal I used to port MOSS from the CDC mainframe to the minicomputer. This was a really interesting exercise and taught me a lot about designing and coding portable software. Very useful later in my career. Oh yeah - since the 4014 was hardwired to the computer, we could run at 19.2 KB. Wow.
By 1980, other individuals began to get involved in the development and maintenance of MOSS, making MOSS the first open source GIS community project. We completed conversion of MOSS to the DG in 1980. During the same period, Harry and George converted WAMS from the HP 3000 box on which the application was originally developed to the DG 330. They completed this work in late 1980 or early 1981. As part of that contracted project, they had to enhance WAMS to allow interactive digitization from either an analytical stereo plotter (the APPS=IV) or from an Altec digitizing table. My part of that project was to design and write a WAMS (soon called AMS) to MOSS translator. This was a real education as I had to learn the physical AMS data structure and was my first introduction to photogrammetry.
A presentation on the early GIS years in Fort Collins can be viewed here.
I joined Autometric in April of 1980. My initial activities were to provide support to the existing MOSS user base as well as to support technology transfer services of implementing MOSS in various BLM and US Forest Service offices. I was also responsible for growing the Autometric office in Fort Collins.
In 1984, I became involved in a new Autometric contract with the US Engineering Topographic Laboratory ( ETL) called TAWS. As part of this project, we all had to learn UNIX. Also in 1984, the United States Geological Survey (USGS) contracted with Autometric to capture requirements and develop a full design document for migrating MOSS from a full polygon data structure approach to a full use of a topological data structure. The MOSS arc/node and DBMS design document: Final draft (https://archive.org/details/mossarcnodedbmsd00reed) report was completed and delivered in early 1985. There are still copies of this report available.
In 1985, the Autometric staff in Fort Collins began discussing the development of a new generation GIS. This new GIS would be UNIX based and would incorporate all of the lessons learned from previous GIS work as well as research done for TAWS, the USGS MOSS Topological Data study, and for the DMA Mark 90 proposal development activity. After numerous discussions, we developed a basic plan and spoke with the execs at Autometric. They said fine, but do the work on your own time. Through 1985, Rob Starling, Jim Kramer, Randy Frosh, John Davidson, Harry Niedzwiadek and I worked on the design of the new UNIX based GIS. We made numerous design decisions, such as the design of a topologically structured spatial database, projection independence, on the fly coordinate reference system transformations, the use and maintenance of metadata, units independence, separation of data from presentation, the use of R-trees and B-trees for rapid indexing, and integrated vector/raster processing. The initial DeltaMap R-Tree software was coded in Fortran 77 by Bruce Blackwell (Autometric and later Oracle Spatial). Bruce's work is described here (Auto-Carto 8, 1987).
In the 1984 time frame, Autometric entered into a contract with the Engineering Topographic Laboratory (ETL) to work on a project called the Terrain Analysis Workstation (TAWS). TAWS, the Terrain Analysis Workstation, was developed with the U.S. Army Engineering Topological Laboratory (ETL). TAWS, a pioneering effort in using the DMA Digital Terrain Elevation Data (DTED), provided capabilities for generating perspective views of terrain, visibility plots, and mobility information.
A key component of TAWS was a GIS. The US ETL decided that the AMS/MOSS GIS software should be posted to the TAWS environment. A key aspect of this work was the hardware platform: HP 9000 series 550 series work stations running the HP-UX operating system. HP-UX for the HP 9000 500 computers was the first commercial UNIX supporting a multi-processor, multi-user system in the early 1980s. This work may be the first known example of a full interactive GIS capability being implemented and deployed in the UNIX environment.
More specifically, TAWS was a prototype of the Digital Topographic Support System (DTSS). TAWS includes four major components: data base development, terrain analysis and product generation, intervisibility analysis and product generation, and environmental effects software. The data base development was a terrain analysis data base which typically includes factor overlays covering soil, drainage, transportation, vegetation, slope and obstacles