I may be generalizing too much, but I might sum up this post as: “uniformity breeds resiliency”—and while I get what the author wants to get across, if everyone had their setup, no one would be able to offer better font or icon defaults as they would have never tinkered.
I agree that contributing better defaults upstream definitely scales better than maintaining one's own customized setup. Of course, those in control of upstream don't always agree on what is "better" or what is a priority.
why would I care about scaling my dotfiles, my zsh custom prompt, or my icon colors?<p>That those things don't scale is hardly news.<p>yak-shaving is fun and an experience in itself. not every activity needs to scale.<p>I just wrote a netfilter_queue[1] based DNS filter in C[2] to extract IDN/punicode domains from the kernel queue and drops the packet.<p>This is only useful to myself and maybe a handful of those who: run a local resolver, disable DoH, and hate punicode/IDN with enough passion to actually want to block it. But it has probably no use in the wider world because it only adds complexity for them. According to this post, now I should feel bad because it doesn't scale? What a strange way of looking at things.<p>[1] <a href="https://netfilter.org/projects/libnetfilter_queue/doxygen/html/index.html" rel="nofollow">https://netfilter.org/projects/libnetfilter_queue/doxygen/ht...</a>)