Stop spending engineering effort solving problems you don’t have
Julian Friedman

I think this is easier said than done, especially for the masses of first, second, or third timers. Working on your first django app? Eventually you’ll google “how to deploy django to the web”, which will lead you to documentation about uWSGI, which will lead you to documentation about apache/nginx/gunicorn. Most people know apache, so atleast theyre back to familiar nomenclature at this point. And they have an intuition about things like ec2, or digital ocean, linode, etc. If theyre (un)lucky enough, they’ll try to containerize their new service. They’ll google “what is a git readonly key” (again, if theyre lucky). At no point will they google “advice from a devops wizard on how the layman should deploy things” — a question to which every wizard has their own answer.

It’s not that I disagree with you (in fact I can only dream of how much simpler your suggestions might’ve made my first few patchwork web platforms), but I think your advice here may be falling upon deaf ears. We need to be steering people down the right path in a more organic way (built into the documentation and stack overflow answer that guide the next generation of web developers), do a better job setting expectations to people playing around with development-mode servers on their laptops (if you’ve got a PoC running locally, you’re still in for an adventure).

Thanks for starting a much needed conversation, but I think this article is pointing the finger at today’s developers instead of at the generation that came before them who now build the infrastructure and frameworks being used.

Show your support

Clapping shows how much you appreciated Bill Smartt’s story.