Nothing being done at all means the system continues working as it is. Something being done incompetently means more people die.

You need to be willing to take risks or nothing will get done. It's okay to accept that people will die. People die every second of every day.

“Willing to take risks” does not imply “let incompetent buffoons manage a major change to air traffic control.”

> I would rather things be done "incompetently" than nothing being done at all.

Engineers of today aren't better than engineers of yesterday, they just have access to new technology (and new footguns).

I'd rather have a well-designed legacy system that does the job well, than an incompetently designed "modernized" system that doesn't.

> Engineers of today aren't better than engineers of yesterday

Our ability to wield technology has gone up, while our ability to engineer and understand that technology (as a whole) has gone down.

Age should be the most suspect of arguments against the design of a system. Enumerate specifically its flaws and not just that it is old. System design doesn't wear out, it is abstract, like mathematics.

Every time I'm in a place of business with shiny computers, I'm reminded of the difference in efficiency between the newest systems and the 'green screen' terminals and other similar tech that was everywhere in 1990.

Most of those were extremely optimized for their specific purposes - for instance, screens were carefully designed to fit all the necessary information to complete a task without any scrolling. They had extremely low latency despite all the hardware being potato-grade compared to what we have today.

When latency existed, simple things like the keyboard buffer allowed the skilled user to get ahead of the machine and move onto another task, while the processing caught up.

Today, it's all a webapp with a ton of React, probably 5,000 npm modules, and you get to watch skeletal loading rectangles pulsating after every click (or branded animations on the "customer-facing" flavor). Instead of having muscle memory of hands on a keyboard (usually a specialized one which gave a dedicated key to every important function), the user must move their arm to find touch targets placed at arbitrary places across a 24" touchscreen, and often hunting for functions inside ••• menus.

Obviously, no usability testing has been done on any of these. Or if it has, the only thing tested was "Can a 17-year-old with zero skills ploddingly complete a transaction (the happy path only!), after only 15 minutes of training?"

So yeah, I never listen to any "newer = better" takes.

I worked on a lot of those "green screen" (or "orange screen") apps.

Limited to keyboard and 24x80 screen, but yes, the apps were lean & responsive -- sub-second response time, and keyboard muscle memory. And some shops I worked in capped your DB calls to <25 per transaction. In stark contrast to some of the Laravel apps I've worked in recent years where the SQL logs showed me 300-400+ DB calls on a given request :(

Your comment reads like poetry.

what an absolutely awful take.

And what an absolutely awful summarization of your argument. The take of let's not change anything ever and trust the process won't work long term.

Why? Do you want more people to die? The only way to migrate a system like this safely is to do it competently and slowly.

Yes it's a risk I'm willing to take. But because of redundancy it will be 0.

I don't. Incompetence costs lives. Using old tech that has worked for decades costs a couple dollars extra to maintain for a system worth millions.

You can justify any waste with that argument. Saying things will cost lives is just a way to interject bureaucracy to slow down a project. You are the enemy of progress even if you don't realize it.

An enemy of progress for not wanting some halfbaked replacement for tech that has worked for decades? Yeah okay.

If it ain't broke, don't "fix" it.