I've never heard of this Keith Geiren, based on this one article he doesn't seem to be a completely frothing at the mouth/privatize everything/typical postmedia opinion writer.

I am far from a technologically-savvy person — certain buttons on my 10-year-old TV remote still confound me — but it seems to me there is a real question here about how much of this dysfunction was inevitable and how much could have been avoided with better planning.

Honestly, having been impacted by processes like these and been involved in planning and building portions of smaller-scale rebuilds... It's basically inevitable. Legacy systems often aren't designed in ways that make it clean or predictable to integrate with other software or migrate to a new platform, and with 70 legacy government systems I would be impressed if they only ended up at double the initial predictions.

Could it be planned better? Sure, but you can easily spend so much time in planning that the goalposts end up moving by the time you start working. With these sorts of projects it's like being handed a ball of tangled cables: yeah, there are strategies to untangle things nicely, but just tugging at it until you get an idea of where the worst knots are is usually just as fast.

So, not a nightmare but complex project is complex?