30 Years in Tech: Castles in the Sand
What do pyramids, the Eiffel tower, the Chinese wall and the Hoover dam have in common? They’re all man-made constructions that have been built to last. And even if they’re not, like the Eiffel tower: they stood the test of time. My average project in the past 3 decades? Not so much. The Lotus-123 ‘data warehouse’ I built 30 years ago is long gone, and so are most of my other contributions to the world. The same goes for many of the organizations or departments I worked for. In fact, my first employer after I graduated, BSO, merged with Philips applications services 18 months after I signed my contract. This was pretty much the end of the organization as we knew it; the structure and especially corporate culture changed dramatically, causing many people (including me) to leave.
From my first workday at BSO in January 1993 until I left in October 1994 I worked on exactly one project: custom development of an ERP like system for a postal distribution organization. It was fun, I learned a lot, but a year into the job I was eager to move on to another project. Unfortunately, I was deemed ‘indispensable’ so after another 6 months or so, I just quit my job to get off the project. The reason I was hard to replace was both a blessing and a curse: scarce knowledge of the technology, the organization and the processes. Great if you want job security, less fun if you want to move on and expand your horizon. The technology? Borland Paradox, and more specifically: PAL, the Paradox Application Language. The DOS version of course. Yep, we built an entire multi-user ERP system using a PC database and ditto programming language. And although Borland has ceased operations long ago, the Windows version of Paradox apparently lives on as part of the WordPerfect Suite. What’s even more surprising: the DOS version still has an active community!
Fortunately, some of my work did stand the test of time. Which brings me to a first observation in this post: the more technology involved, the shorter the life span of your efforts. What also could be true is that maybe commercial organizations change technology more often than government, or are forced to do so due to mergers or acquisitions. Point in case: 25 years ago I was the project lead to implement a new methodology of funding primary education in the Netherlands. It was more focused on business process and communication changes; the impact on IT was fairly limited. And, as far as I can tell the resulting system is still in effect.
The second project that outlived my own involvement is one I’m particularly proud of. It’s a database system for scientific research of pharmaceutical prescriptions and effectiveness on patients for the Santeon Group, a cooperation of 7 top-clinical hospitals in the Netherlands. The challenge was to design a system where patient data could be safely and anonymously shared for research purposes, but when research showed alarming results for one or more patients, the hospital (as data owner) should be able to identify which patients it concerned. Other requirements were the ability to easily create subsets of patients who had given consent for a specific research project, but not for others. Fortunately I was part of a dream team at Antonius Hospital back then (which certainly deserves its own separate post). Based on the clinician’s requirements we built a prototype based on Pentaho Mondrian, the open source ROLAP engine which was well received. After I left the team early 2013 the system was successfully implemented using the same open source technology and is still in active use today!
Those are the exceptions though. Most of the projects and systems I worked on have washed away like the sea washes away castles in the sand. Which brings me to the final lesson of this post: smile, and relax. There’s more to life than worry about systems that no one will use anymore 5 years from now.
