Post GIS

Early in my career, I was interviewing for a job with a large, three-letter, consulting firm. I was going to be the “GIS guy” on the team. The interview was wide ranging and went well. We eventually got around to the topic of dynamic maps on web sites. To place this in the proper technology context, this was 1998.

As it happened, I had some experience with web mapping as I had just implemented ArcView IMS for my soon-to-be-former employer. Avenue and HTML, FTW!

The hiring manager, being a mainline IT guy and a tech geek at heart, began to postulate how they must work. In this mind, there was a database full of small JPEG or GIF images that you queried and sent to the browser, based on the user’s current map extent. This, of course, was preposterous to a “GIS guy” like me.

I began to educate him on how GIS renders maps dynamically, as well as other important concepts like map projections, storage requirements, and the like. Armed with this new knowledge, the hiring manager came to two conclusions: 1) I was the guy for the job and 2) His original concept was obviously off the mark. It’s a good thing I was there to save him from himself.

Those of us who were around on February 8, 2005 and beyond will recognize that I talked that manager out of map tiles.

The reason I relay this story now, and the reason I chose PostGIS Day to do it, is because the geospatial world is moving ever more rapidly into its post-GIS (as in “after GIS”) phase. Map tiles (ala Google Maps) was just the opening salvo in a decade of progress in which the web, looking at traditional GIS, said “no, thank you.” The result has been a period of innovation that has brought geospatial information and methods to the web on the web’s terms. GIS has had to adjust.

That’s not to say that geography hasn’t had something to say. The recent rediscovery of the value of map projections, most obvious in the support of projections in D3, indicates that the discipline of geography might have one or two good ideas to consider.

But the trend has been clear: geography, yes; GIS, no. Support for spatial data is now taken for granted in databases and it is now fashionable to state that data science is eating GIS.

GIS is, as it should be, fading into the overall information landscape and becoming less distinguishable as a distinct entity. The value of what it does is much greater than the value of what it is.

So how does this relate to my map tiles story? While the past decade has seen tremendous progress in the world of geospatial technology, it was preceded by a couple of decades of building up a walled garden in which GIS was seen as special and exotic. Given the severe limitations of computing power over much of the history of GIS, as well as the formidable knowledge required to be a good geographer, that was probably justified.

But, architecturally, the web and the “cloud” and data science have caused us to rip apart the old monolithic software tools so that we can take just the bits we need for the job at hand. This approach makes GIS more useful, more accessible, but no less powerful.

It’s important to not get stuck in old modes of thinking with regard to GIS. By the time I loaded Google Maps for the first time, I already had a few more years of working with geospatial technologies under my belt and was starting to rethink my approach to implementing them. I was about a year away from discovering open-source geospatial tools, of which PostGIS was my gateway drug.

But seeing those map tiles render the first time made me instantly recall that conversation from a few years before. The fact that my hiring manager and Google had one thing in common was not lost on me: neither had a traditional GIS background. Because of this, they were free to think of the problem on different terms. I realized that I couldn’t let my experience with GIS become a limiting factor and I vowed never to get caught flatfooted again.