alt Hacker News> It has become a big pet peeve of mine when people treat "workarounds" like "solutions" to problems. I have certainly done this in the past, so I'm not excluding myself from it, but I try pretty hard not to do that now.
My favorite is this pattern that occurs in every big backend job-running script in every place I've worked: the success paths spams the log with expected errors. Something tries to connect to something else on start?
FATAL ERROR: COULD NOT CONNECT
debug: retrying... (1/3)
FATAL ERROR: COULD NOT CONNECT
debug: retrying... (2/3)
Service connected!
Startup succeeded
I used to work at a very large fruit company. I won't say the name because I don't want you to Think Different about them (and because I don't want as much correlation data about me on this account).
On the first day, one of the first things they had me do was set up email filters with an elaborate home-built email filtering system. The reason for this was because I would get thousands upon thousands of emails per day (not hyperbole), most of which were irrelevant to me, and if you didn't have fairly fine-grained and elaborate filtering your important emails would certainly be lost, often within minutes.
The solution to this problem, of course, would be to stop sending so many emails, or have better control over who was getting the emails, but instead the onus was put on everyone downstream to figure out which emails were spam (and to get yelled at if we didn't respond to an email because an important one got caught in the mix).
I complained about this a few times, and people's responses would always respond about how filters could solve this problem, and it always annoyed me. If someone is dumping a metric ton of trash in my backyard every day on top of my Amazon packages, the solution is not to figure out an optimal way to categorize and sort the trash to best differentiate it from my packages, the solution is to get that person to stop dumping garbage on me.
They were unimpressed by this reasoning.