Back in 2000, it was the wildwest. You actually did run tests on production and editing production code was not as insane of an idea as it is today.
I don't think anyone (at least no one I worked for at the time) ran staging or dev servers.
It was always stupid, yes. But back then we didn't have the tools and testing suites that we have today. CI/CD setups didn't exist. Git wasn't even built until 2005. The only version control solution was SVN at the time, which was released in 2001. But it was clunky and immature.
Back in 2000, launching a site update meant someone would log into the server via FTP, drag the files over, and try to "be careful" while they did it. Using passwords that were written on a sticky note, stuck to the CRT monitor's screen (password managers weren't a thing).
It is easy to forget how immature and primitive the world of web development was at this time.
Also, the internet was still so new, that every C-level executive had built their careers by running businesses in the 70s, 80s, and 90s. Back before you built websites or relied on them for any significant impact on your bottom line. So to tell an executive from that era that your website broke when you poked at it, their solution would be to stop poking at it. Security wasn't really a major concern like today, and having a website was still mostly a novelty in the eyes of most executives.
> Back in 2000, launching a site update meant someone would log into the server via FTP, drag the files over, and try to "be careful" while they did it.
Heck, I remember just compiling Java classes on the server machine itself, copying them to a production directory, and restarting the app server (tomcat IIRC). Source control was the sysadmin running a nightly backup of source directories
> The only version control solution was SVN at the time
Pre SVN there was CVS, and before that was RCS. Now these didn't work the way we think about git today, but they did allow you to roll back to a known good state with some futzing about.
I don't think anyone (at least no one I worked for at the time) ran staging or dev servers.
It was always stupid, yes. But back then we didn't have the tools and testing suites that we have today. CI/CD setups didn't exist. Git wasn't even built until 2005. The only version control solution was SVN at the time, which was released in 2001. But it was clunky and immature.
Back in 2000, launching a site update meant someone would log into the server via FTP, drag the files over, and try to "be careful" while they did it. Using passwords that were written on a sticky note, stuck to the CRT monitor's screen (password managers weren't a thing).
It is easy to forget how immature and primitive the world of web development was at this time.
Also, the internet was still so new, that every C-level executive had built their careers by running businesses in the 70s, 80s, and 90s. Back before you built websites or relied on them for any significant impact on your bottom line. So to tell an executive from that era that your website broke when you poked at it, their solution would be to stop poking at it. Security wasn't really a major concern like today, and having a website was still mostly a novelty in the eyes of most executives.