>> you lay out a huge specification that would fully work through all of the complexity in advance, then build it.

> This has never happened and never will. You simply are not omniscient. Even if you're smart enough to figure everything out the requirements will change underneath you.

I am one of those "battle-scarred twenty-year+ vets" mentioned in the article, currently working on a large project for a multinational company that requires everything to be specified up-front, planned on JIRA, estimates provided and Gantt charts setup before they even sign the contract for the next milestone.

I've worked on this project for 18 months, and I can count on zero hands the times a milestone hasn't gone off the rails due to unforeseen problems, last-minute changes and incomplete specifications. It has been an growing headache for the engineers that have to deliver within these rigid structures, and it's now got to the point that management itself has noticed and is trying to convince the big bosses we need a more agile and iterative approach.

Anyone who claims upfront specs are the solution to all the complexity of software either has no real world experience, or is so far removed from actual engineering they just don't know what they're talking about.

Working on a project for 18 months doesn't give you enough insight into it to know what is good or now about it. You need several more year before you can usefully figure out what changes will help get you to make milestones. (other than trivially obvious things, which might be the low handing fruit - but sometimes they are the better way to do things but the real problem makes that stand out instead).

Nothing will get you to hit every milestone. However you can make progress if you have years of experience in that project and the company is willing to invest in the needed time to make things better (they rarely are)

Agreed. Since this blog has posts since 2007 I can only think that the author is in the "so far removed" group.