Comment by Attrecomet
1 year ago
What is the alternative, though? This is an honest question, I really want to know: How would the "this is horrible, how can you deign to work that way?" crowd coordinate thousands of people on a project to create something that is bigger than what 5–20 people can create?
Because most of the answers I see here gloss over that part, or strongly imply that engineers will always decide better than business people what should be released. And I can sympathize, especially if you are in an MBA-led org, but I am also certain that if you think you know perfectly what the enterprise or customer needs, and anyone opposing you is a Pointy-Haired Boss, that you are most probably the idiot in that case: 90% of the time a single dev will NOT have better business intelligence than everyone else.
> What is the alternative, though? This is an honest question, I really want to know: How would the "this is horrible, how can you deign to work that way?" crowd coordinate thousands of people on a project to create something that is bigger than what 5–20 people can create?
Nobody needs to be co-ordinating thousands of people. 5–20 people can create Instagram. The entire problem in these companies is that leadership is so out of touch they cannot differentiate between a checklist and a product, and empire-building is their proxy for value. The solution is to change the leadership, but it is usually too late in large orgs (the new leadership has to be brought in somehow from somewhere, and that will be done the same way the current leadership happened).
So the real solution is for those who care to go elsewhere, out-compete, and out-succeed. Then quit after acquisition, if such a thing happens.
"Just never develop software in a medium or large company" is a take, I'll give you that. I'm pretty sure that the vast majority of software developed is vastly more productive than Instagram, and tied to real world processes that need coordination. Your 5 people team will do fuck all to program the control software for a crane arm activator, because they'll never get even close, and if they were actually given access, you would have to coordinate with hardware people and actual engineers (the ones with detailed plans and calculations before building anything, I mean, not us software "engineers"), you would have to figure out what the construction companies using it actually need from your software. A similar story could be told for most any area of software development - it generally is tied to other areas, and unicorn examples of how wonderfully easy the world is if you develop a time wasting app for phones is not applicable to the majority of us.
I totally agree with you last point. Not the least for the egoistical reason of higher chance of better products for me, that way :)