Porting a venerable library on vibes. What started as a quick WebPerl success spiraled into weeks of failed attempts, mangled ASCII diagrams, and a hard lesson about respecting the craft embedded in battle-tested code.
Chewing up decades of careful work and spitting it out with a gaggle of coding agents is disrespectful to the craft.
I spent weeks casually trying to replicate what took years to build. My inability to assess the complexity of the source material was matched by the inability of the models to understand what it was generating.
I absolutely agree that attempting to regenerate 30k lines of code into another language is incredibly overconfident. I’m sure you could probably generate a few hundred lines of code that function as expected. To generate that much code, I expect constant maintenance of the output and a moderate up deep understanding of both the old and new code.
This was the bias I entered with.
Crappy prompts and overconfidence will produce poor results.
I absolutely agree that attempting to regenerate 30k lines of code into another language is incredibly overconfident. I’m sure you could probably generate a few hundred lines of code that function as expected. To generate that much code, I expect constant maintenance of the output and a moderate up deep understanding of both the old and new code.