Insularity and Progress

This article was first written for Earth Observation Magazine in 2004. It is still accessible but only on the web archive, so I thought I would take a copy.

Insularity and Progress
Simon Greener

Original Article

My 20 years in IT and GIS range from general IT on mainframes to GIS consulting in business and academia to GIS Manager at a large Australian company. The thread linking these experiences for me is how important databases are in data management. When I entered GIS, we lacked generic database products for managing spatial data. Databases 101 taught me the values databases bring to IS management: first, the model, embodied in the catalogs at the heart of commercial databases; second, the logical separation of application from implementation. We had to wait until the 1990s and early 2000s for databases to deliver these accessibly and affordably, via object-relational products like Oracle Spatial, Informix IDS, and PostGIS.

Unfortunately, uptake of these products within GIS has been slow. I want to suggest a possible non-financial reason for this.

Sometimes I think I must have two heads. In building an enterprise-class, flexible, stable, open, extensible, and cost-effective GIS infrastructure for my company, I chose a commercial database in preference to “enterprise” products from a large player in GIS. When I try to explain why to my peers, I use arguments drawn from my computing science training, promoting management principles informed by good theory and proven in practice.

Most of my peers walk away unconvinced. This may be because I don’t use lucid, digestible arguments that draw on my peers’ experience. Or it may be because of something my colleagues share with the many students I have lectured and mentored a problem endemic enough in GIS to warrant this Soapbox.

I have been concerned awhile that some higher education institutions unintentionally promote the notion that “GIS = skill with vendors’ products.” Now, I do not contest the quality of these products, or the motives of vendors in gifting institutions with their software. But the consequence is that when it comes time to implement geographically aware solutions, the learned assumption is: if it is geographic, you must use a GIS; or, worse, you must use a GIS from vendor X.

I often think the fascination of GIS vendors with physical file formats harks back to the pre-database computing of the 1960s: so many seem to subscribe to a version of the old adage “nobody ever got fired for buying IBM.” The only thing that’s changed is the vendor’s name.

Is our industry too insular? Pontifications about how “special” Geographic Information is might be good for morale, but hardly for sound decision-making. Is our professional education too insular? Are we so familiar with a few vendors’ products that we won’t step outside the comfort zone to investigate “non-proprietary” solutions?

The answers are not simple, but surely one of our problems is an industry Zeitgeist that fosters a “cradle to grave” approach to software acquisition. Thus, organizations invest their key GIS infrastructure in single-vendor stovepipe products. While this investment makes life easy for GIS managers, it makes it hard to convince them of the value of non-traditional products like databases. Truth is, such “safe” investments make information systems restrictive, inflexible, and costly. They also delay the uptake and development of computing technologies that lie at the heart of the latest “open systems” revolution.

Our industry’s response to this revolution has been to convince itself that the latest incarnation of this or that GIS will work in an open computing environment. Father knows best: the familiar products really can offer choice in developing fully spatial IS.

I beg to differ.

New Observations.

I still think the article makes a good point about vendors, education and industry development.

What I have noted since that time, is that things may have gotten worse, yet at the same time some aspects have improved enormously: the continued growth of FME, qGIS, R and geoServer.

What is not good for the industry is vendor monoculture.

The infiltration of some GIS software vendors into higher education is pretty well complete. (Great if you are that vendor; hard luck if not.) Courses have become more focused on industry requirements though skill with day to day software packages. Or at least what they “think: industry wants. Are graduates capable of using software at the new job, that they didn’t use at University? Is it a reasonable thing to then undermine the existing vendor in their new work environment to that it is replaced with only that software that they are familiar with?

This affects open source uptake as well as commercial vendor update.

The vendors have cleverly changed their marketing strategies to include offering software to Government Departments and other businesses for free for a fixed amount of time before starting to charge. While there should be questions asked about Government Departments accepting such offers without a following an open, transparent and competitive tender, one has to ask if this is good for anyone except the profits made by the vendors?

This sort of sales and marketing approach works because of the increasing corporatism of society. Decisions are made by big political parties, big government, big environment, big welfare, and big businesses. One aspect that I suspect may be going on is sociological, in that decisions are made more by seeing what others are doing than determining what is best for the organisation. “Best Practice” becomes a euphemism for single vendor purchase. Perhaps one example could be the preeminent position of SAP within certain markets in Australia and elsewhere.

When market power created by the sorts of practices and environments mentioned in this article, is used by software vendors to close out competitors, we are on that point where Capitalism gives way to Corporatism.

I don’t think we want to end up there.