When I first studied cartography back in the technological stone age — 1979 — I drew maps by hand. All research and data collected for maps was either done in the library or out in the field, and the results were compiled and tabulated by hand for manual analysis that was eventually drawn on the maps. I even had to calculate the projection I wanted to use and then project the map based on a globe. Cartography was a very time consuming process. By 1980, we were allowed to use the mainframe at the university to analyze some of our data. I wrote programs using Statistical Analysis System (SAS); each line was a punchcard (a typical program had hundreds of punchcards), and the punchcards were fed into a card reader. Then I sat in the computer center drinking coffee and doing homework for great lengths of time until the output of the program came back. If there were errors, I would fix the lines, produce new cards, replace them in the stack of cards, feed the card reader, and wait for the next output with errors or results of the program. Even though I got to use the computer to run the stats, I still had to map the results by hand. By the time I graduated in 1983, GIS was just coming into use at the university.
I was thinking back to those days today as I struggled with getting map servers working on my workstation for testing and development, before I deploy them on our production servers. As I hacked installers of the various components to force them to find dependent programs that had different paths to the applications than the installers expected and beat my head against the desk hacking postgres so I could force-install postgis needed to deal with spatial data, I started to wonder if the world of GIS and web mapping were really advancements over old fashioned cartography. Once the hacking was done, and I had my workstation serving maps on the web, I thought better about doubting GIS’s ability to process and map large data sets in seconds that would take weeks or months to process and map by hand. In reality, the frustration and time spent hacking installers, and fighting programs to get them to work the way I want them to is well worth the trouble for the resulting efficiencies in mapping services we can offer. Although all this efficient technology ends up taking a lot of the art out of mapping.