Early Geospatial Computing at Harvard: Some Old Gisers Try to Remember

Panel Members:

Nicholas Chrisman (GEOIDE, Quebec QC, CN); Geoffrey Dutton (MathWorks, Natick MA, US); Scott Morehouse (ESRI, Redlands CA, US); Allan Schmidt (Concord MA, US); With contributions from Denis White (EPA, Corvallis OR, US)

Geoff Dutton 
On Integrating Spatial Analysis and (Geo)graphics: The Warntzian Legacy

William Warntz hired me to analyze data with him at the Lab in 1967. To my knowledge, he never programmed or even used a computer nor had any opinions about user interfaces or file formats. He understood CDBs to the extent of SYMAP packages, and while he had knowledge of topological data structures, he didn't talk or write about them. The Warntzian paradigm "Geography, Geometry, Graphics" informed perhaps half of what the lab did. Except by implication, it does not encompass other major research themes, such as Topology, Data Structures, Cartographic Data Bases, and User Interfaces. The computational slogan "Raster is faster but vector is correcter" summarizes ongoing frictions in and around the Lab.

• Geographic computational geometry (point pattern and network analysis; mathematically 
  rigorous but empirically unprovable)

• Landscape analysis models (raster approaches such as “map algebra”; theoretically shaky, yet 
  with empirically plausible results)

Bill Warntz believed that these sub-disciplines needed to talk and make themselves useful to one another. It took a long time, but slowly, raster and vector approaches to spatial modeling did sort of connect. Lab people who helped make this happen include:

• Tom Poiker & Co. – Sampling terrain surfaces with triangulated irregular networks
• Scott Morehouse & Co. – ESRI software to convert between triangulated and raster representations
• Denis White – environmental sampling using planar hierarchical tessellations
• Geoff Dutton – Global hierarchical tessellations; GEM (terrain) and QTM (anything)
• Warntz himself – Without his concepts of topology of surfaces, TINs and hydrological modeling might have taken longer to evolve, and GEM/QTM wouldn't have happened.

Scott Morehouse
Lasting Lessons from Working at the Lab

• Building generic software tools for geographers & landscape scientists (simple data model, 
  algebra of tools, design, build, ship, iterate)
• Team-based, modular programming (modular FORTRAN; looking over shoulders, discussing,
      and doing)
• Map definition language (DOT.MAP, Polyps) (the anatomy of a map;  a language for defining 
  "symbology", "map surround elements"; etc.)

Allan Schmidt 
Harvard University's contribution to the evolution of GIS, 1833-2008

• 1833 Despite "barely getting in" to Harvard College, Henry Thoreau maintained above-average grades and studied classical literature, and French, Italian, and German, as well as math, geology, zoology, botany, and natural and intellectual philosophy.
• 1837 Thoreau receives his degree from Harvard College, graduating 19th in a class of 44.
• 1847 LouisAgassiz appointment as Professor of zoology and geology at Harvard proposes theory of glaciation.
• 1847 Henry Thoreau studies surveying, works as a surveyor in Concord MA (1850s) using chain and compass.
• 1850-1860 Henry Thoreau's GIS contribution and Walden Pond-Brister's Spring mistake.
• 1894 USGS mistake.
• 1966 Howard Fisher establishes the Harvard Lab for Computer Graphics
• 1967 Harvard GSD Newsletter announces formation of the Lab.
• 1981 Dutton/Cohen LCG report on the history of the Lab
• 1968-1979 LCG Context newsletters
• 1966-1970 LCG Redbook editions
• 1970’s-1980 LabLog products and services catalog
• 1979-1981 Harvard Library of Computer Graphics
• 1978 Harvard Papers on Geographic Information Systems
• 1980 Harvard Business Review article
• 1995-1999 Niemann Interviews of Chrisman, Morehouse and Schmidt
• 2003 Jon Corson-Rikert LCG web page at Cornell Library
• 2006 Nick Chrisman's book "Charting the Unknown"
• 2007 Harvard CGI legacy
• 2010? Monmonier book re History of Cartography (forthcoming)

Denis White 
Lasting Influences on My Work

Two disparate influences from the lab have prospered in my life at EPA since then. Mike Woldenberg's work on branching structures and tessellations somehow got embedded and has fostered applications in environmental monitoring and fish modeling, among others. Carl Steinitz's alternative futures style of landscape planning has invigorated a broadly applied approach to scenario-based environmental assessments.

Nick Chrisman 
Charting the mostly remembered

• Writing a book about the Lab: issues in first-person history (Privileged access to the documentary records—confessions of a packrat, and friend of other packrats; Memory? Selective accounts to coordinate; It was easier to write the 1966-1971 part)
• Local vs. global: why we thought we didn't need no stinking database
• The history of an artifact: the curious evolution of metadata

As soon as elements of ODYSSEY started working, we needed some way to save just enough to make them work together. The question of what was enough rages to this day… (various GISers will take their habitual positions on the spectrum from simplicity to full disclosure).

• What endures: A round-table