The article says "I think the biggest factor is that any rewrite of an existing codebase is going to yield better results than the original codebase.".
Yeah, sorry, but no, ask some long-term developers about how this often goes.
replies(2):
Yeah, sorry, but no, ask some long-term developers about how this often goes.
I've been a software developer for nearly 2 decades at this point, contributed to several rewrites and oversaw several rewrites of legacy software.
From my experience I can assure you that rewriting a legacy codebase to modern C++ will yield a better and safer codebase overall.
There are multiple factors that contribute to this, such one of which is what I reffer to as "lessons learnt" if you have a stable team of developers maintaining a legacy codebase they will know where the problematic areas are and will be able to avoid re-creating them in a rewrite.
An additonal factor to consider is that a lot of legacy C++ codebases can not be upgraded to use modern language features like smart pointers. The value smart pointers provide in a full rewrite can not be overstated.
Then there's also the factor that is a bit anecdotal which is I find that there are less C++ devs in general as there was 15 years ago, but those that stayed / survived are generally better and more experienced with very few enthusiastic juniors coming in.
I'm sorry you did not enjoy the article though, but thank you for giving it your time and reading it that part I really appreciate.