I found the "finger guns" team and their plan fascinating, and I've definitely seen enthusiasm sweep the field many times.
I work on an infrastructure / foundational engineering team, and it reminds of something I'm increasingly believing to be true:
It's much harder to get headcount/funding for a two-year plan to build/fix an internal tool than to get headcount/funding for a six-month plan to use an external product that ends up taking 3-4 years.
In this situation, I'd say it's much harder to get headcount/funding for a five-year plan to perform a well-engineered migration than it is to get approval for a two-year whizz-bang plan to create an amazing replacement that will actually take 3-6 years and/or actually leave 80% of the legacy code migration unfinished basically forever, leading to pure xkcd 927.
As an addendum, I found the episode fascinating. As someone who has spent about five of the last eight years on a foundational engineering team, involved in _many_ long, grinding migrations where the end goal is clearly a win for the company but not particularly useful for product teams in the short term, I thought the plans Chris outlined showed remarkable maturity and good planning in approach. I also enjoyed hearing a public discussion of the nature of that kind of migration.
I would not, myself, choose to talk quite so negatively about a former employer and colleagues, but I respect and enjoyed the decision of idealistic folks like Chris to be open about the struggles entailed!
I work on an infrastructure / foundational engineering team, and it reminds of something I'm increasingly believing to be true:
It's much harder to get headcount/funding for a two-year plan to build/fix an internal tool than to get headcount/funding for a six-month plan to use an external product that ends up taking 3-4 years.
In this situation, I'd say it's much harder to get headcount/funding for a five-year plan to perform a well-engineered migration than it is to get approval for a two-year whizz-bang plan to create an amazing replacement that will actually take 3-6 years and/or actually leave 80% of the legacy code migration unfinished basically forever, leading to pure xkcd 927.