Sintax design, not implementing a parser.
You're arguing that you might not make up your mind and feel compelled to churn parts of your implementation due to your indecisiveness.
That's not the parser's fault, and I fail to see how anyone avoids this indecisiveness by replacing a proper parser with a bunch of ad-hoc hacks.
> chances are really high that you're gonna want me to do something different.
So what? Just work on it. How many parser generator tools are there? It's as if people have been doing this for decades.
Disagree. The fact that you're working on "a parser" creates the bad project structure that leads you to get hung up on details: a parser whose backend doesn't exist yet inherently has unclear requirements and a lot of bikesheddy decisions to make.
I fail to see your point. OP was commenting on the technical difficulty of actually rolling out a parser, but you're not arguing the technical side. You're arguing the project management sides where "endless bike shedding" results in code churn. That's not the parser's fault, but the person's fault. You're bound to get hung up on details regardless of what part of the project you're working on, because it's not the parser compelling you to get hung up on details.