As my beard starts to grey, I've come to the conclusion that needing to drop a lot of money to retrofit is an infinitely better problem than not having captured a large segment of the market.
As someone who did Perl commercially from 2001-2019, most of my career was spent working at companies who'd done a large Perl build-out some time between 1995 and 2005, made a lot of money and captured a large amount of market share, and were now expensively retrofitting a lot of very, very nasty code that inexperienced programmers had smashed together in a rush.
Much better to be in that position than a company with a beautiful codebase who didn't capture the market
For sure. But I wonder how many are prevented from capturing a larger segment of the market because of poor design decisions re more fundamental primitives of the product. I suppose the worst case scenario here is a complete rebuild and migration.