3 ms·

False dichotomy. "If you deploy this code as is to production there is a high risk of bug X which I think is unacceptable risk. We had a similar bug 3 months ag

by pyfon 1y ago

False dichotomy. "If you deploy this code as is to production there is a high risk of bug X which I think is unacceptable risk. We had a similar bug 3 months ago.". Click the request changes button. Simple. If someone keeps making mistakes, talk to their manager maybe. .

notepad0x90 1y ago
That problem is already caught, how do you make sure they know and understand that mistakes like that cause harm and it was a result of their negligent attitude? Fixing that mistake isn't the goal, making sure they exercise more care and understand the gravity of the situation is.
- viraptor 1y ago
  "making sure they exercise more care" is quite a vague concept. If things are important there will be checks and extra safety nets and people will learn over time. If the only thing stopping harm is people "caring more", then you have a ticking time bomb on your hands anyway. And yeah, you can just tell someone "this is very important and serious" to convey it's very important and serious. Adding asshole phrasing on top doesn't change much.
  - notepad0x90 1y ago
    it doesn't work that way in reality. You can monitor/check for bugs you can anticipate, but some bugs fall through the cracks and you have RCE's and outages causing millions of dollars. There is a limit both in terms of imagination and available resources on what you can anticipate and prepare for. There will never be a world where humans won't have to catch corner cases and bugs that would cause severe harm.
    - viraptor 1y ago
      I'm not saying the checks and safety nets have to be automatic. QA, reviews, design sessions exist for a reason.
      - notepad0x90 1y ago
        I understand, considering those as well and this goes beyond software dev as well. All those things exist to catch mistakes, they don't exist to prevent them. There is no replacement for humans needing to be careful. This isn't even a tech issue, in every industry where safety is required there are many checks, balances, tests. But you still need to not screw up. I screw up, everyone does, but being responsible and not repeating mistakes, that is what I want to communicate better without being an ass.
- pyfon 1y ago
  The manager needs to judge if it is a one off oversight, a skill issue or an attitude issue. I think it is better to talk to everyone civilly and respectfully regardless and then use fair processes to deal with these issues. Broadly you have 2 possible issues: 1. People make mistakes and you need processes like code reviews and tests etc. to catch them. No one does 100% perfection. Avoid blame. 2. Someone is not performing well enough. You deal with 1 as a team through continuous improvement. And 2 is usually manager and team member through feedback and perhaps performance improvements. It should be utterly respectful at all stages.

by pyfon 1y ago

notepad0x90 1y ago

That problem is already caught, how do you make sure they know and understand that mistakes like that cause harm and it was a result of their negligent attitude? Fixing that mistake isn't the goal, making sure they exercise more care and understand the gravity of the situation is.

viraptor 1y ago
"making sure they exercise more care" is quite a vague concept. If things are important there will be checks and extra safety nets and people will learn over time. If the only thing stopping harm is people "caring more", then you have a ticking time bomb on your hands anyway. And yeah, you can just tell someone "this is very important and serious" to convey it's very important and serious. Adding asshole phrasing on top doesn't change much.
- notepad0x90 1y ago
  it doesn't work that way in reality. You can monitor/check for bugs you can anticipate, but some bugs fall through the cracks and you have RCE's and outages causing millions of dollars. There is a limit both in terms of imagination and available resources on what you can anticipate and prepare for. There will never be a world where humans won't have to catch corner cases and bugs that would cause severe harm.
  - viraptor 1y ago
    I'm not saying the checks and safety nets have to be automatic. QA, reviews, design sessions exist for a reason.
    - notepad0x90 1y ago
      I understand, considering those as well and this goes beyond software dev as well. All those things exist to catch mistakes, they don't exist to prevent them. There is no replacement for humans needing to be careful. This isn't even a tech issue, in every industry where safety is required there are many checks, balances, tests. But you still need to not screw up. I screw up, everyone does, but being responsible and not repeating mistakes, that is what I want to communicate better without being an ass.
pyfon 1y ago
The manager needs to judge if it is a one off oversight, a skill issue or an attitude issue. I think it is better to talk to everyone civilly and respectfully regardless and then use fair processes to deal with these issues. Broadly you have 2 possible issues: 1. People make mistakes and you need processes like code reviews and tests etc. to catch them. No one does 100% perfection. Avoid blame. 2. Someone is not performing well enough. You deal with 1 as a team through continuous improvement. And 2 is usually manager and team member through feedback and perhaps performance improvements. It should be utterly respectful at all stages.