Thanks, but I don't think it was really anything to congratulate, because I was literally starting from a blank slate looking for the combination of tool and documentation that would make my life easiest. And its entirely possible that I took the harder path, but from the available information at the time it seemed like the right choice, and having learned the tool I think it has saved me effort over the years. I'm just contributing my experience to the discussion where the goal is probably to determine the path with the lowest cognitive overhead for a new module author.

I wrote a script which does that for me

and when I started I wouldn't have known what a script like that would need to do. It takes time to learn the structure of a package and read about all the config options for MakeMaker to find out which parts are already solved and which need solved with additional scripting.

I do agree that its unfortunate that dzil is based on such a large dependency tree. On the upside, all you need is cpanm Dist::Zilla; dzil authordeps | cpanm to get started, and you can use perlbrew to make a self-contained perl to use for the pull request challenge, and then delete it all when you're done.