That’s a tagline I certainly never expected to write, although the notion that we’d visit these issues again has occurred to me periodically over the past 25 years. Stepping into the “way back machine,” I have many memories, some of them fond, of the last few years of the 1990s. We were in a mad rush to eradicate the issues associated with the Y2K “problem,” something that resulted from a deliberate decision at the dawn of the computer era to represent years as two-digit code. When computing power and storage were high-priced commodities, this made sense, and no one thought that the systems of the day would persist for 30-40 years. I hesitate to call it a “bug” since it was actually a logical, planned decision which just happened to have some spectacularly unfortunate unintended consequences.
To fix this, many old systems were replaced or remediated in order to accommodate date fields that could reflect both century and year. Field expansion was a great, permanent way to fix the issues; it also happened to be a pretty expensive cure, so it wasn’t universally adopted. For many applications that couldn’t possibly survive for another 20-30 years, windowing became an effective—if temporary—solution. Logic applied against dates would interpret a range of 00-20 as the 21st century; everything else as the 20th. Problem solved, time to move on.
“Those Who Cannot Remember the Past Are Condemned to Repeat It”
Which brings us to today. Having reached the end of the logic window, some companies are experiencing another set of unintended consequences. A recent ZDNet article highlighted the problem, noting an example of parking meters in NYC waking up this year thinking it was New Year’s Day 1920. They are hardly alone. A large insurer Novarica recently met with had the experience of producing its first round of bills this year also thinking nostalgically about the end of the Woodrow Wilson era. Not a fatal flaw, but certainly not good.
A notable number of technology organizations are addressing the issue now. Of course, 2020 isn’t the only year that may be part of a “windowing” event, since 2030 was also a target year. In addition, there’s a UNIX “flaw” with a similar origin that will manifest itself in 2038. While it’s not nearly as serious as Y2K was, it is nevertheless something to be aware of.
“The Future Depends on What You Do Today”
All this points to the importance of being aware that there will always be consequences to the decisions we make in a world where there are very real constraint issues to be managed. Every trade-off produces a benefit, but it might also introduce a new issue to be managed later. Examples of this are easy to find; saving money on disaster recovery sites was one of my personal challenges, since this sounded great right up to the point when we had to exercise those plans. Explaining recovery time objective (RTO) and recovery point objective (RPO) concepts while in the middle of a true disaster is a painful and pointless effort, but learning lessons from the experience? Invaluable.
Most organizations won’t have a mark left on them from the “Y2K+20” problem, and there’s plenty of time to address 2038. But how well-formulated are the current IT plans developed by organizations, and do they effectively have room for rapid changes in technology, sourcing models, distribution channels, and customer demographics? In a world that is moving fast, having an effective roadmap for the future accompanied by a series of well-articulated guiding principles is critical to the type of IT/business unit collaboration required to maintain or improve competitive positioning. It can also avoid some really expensive detours. More details on how Novarica helps carriers with these IT strategies can be found here.
Add new comment