Yeah, I don’t think I agree with you at all. Software development and operation are vastly different jobs. Packaging is yet a different story. Maintainers need different things than developers. Handling dependencies is a chore, and you need lots of them if your product speaks dozens of protocols and can interconnect with thousands of devices, each with their own quirks… All the people have something in mind. They already pay attention to deployment and support several methods. Sure it’s not the method you have in mind. But the world doesn’t specifically revolve around you. There are other factors at play. And sure. It’d be awesome if we solved software packaging, dependency hell, the supply chain of larger projects and everything. It’s just not easy. And reality has quite some limitations. It’s just… fighting reality doesn’t get you anywhere. Sometimes we have to make ends meet with imperfect solutions. Or you just live without a smart home. Or use a different software stack. I mean there is FHEM and some other projects.
And with that said, there is some merit to what you’re saying. Software should be designed with usage in mind. It’s just not easy and there are contradicting requirements. Either someone puts in all the effort to cater for your specific use-case… Or they don’t.
They already pay attention to deployment and support several methods. Sure it’s not the method you have in mind. But the world doesn’t specifically revolve around you.
It’s not my method. Writing software with distributions in mind is the standard in free software development.
It’s just not easy.
Indeed. That’s why many engineers don’t bother. Especially poor engineers.
Yes, often projects which are engineered without distros in mind. Which is to say, engineered poorly.
Yeah, I don’t think I agree with you at all. Software development and operation are vastly different jobs. Packaging is yet a different story. Maintainers need different things than developers. Handling dependencies is a chore, and you need lots of them if your product speaks dozens of protocols and can interconnect with thousands of devices, each with their own quirks… All the people have something in mind. They already pay attention to deployment and support several methods. Sure it’s not the method you have in mind. But the world doesn’t specifically revolve around you. There are other factors at play. And sure. It’d be awesome if we solved software packaging, dependency hell, the supply chain of larger projects and everything. It’s just not easy. And reality has quite some limitations. It’s just… fighting reality doesn’t get you anywhere. Sometimes we have to make ends meet with imperfect solutions. Or you just live without a smart home. Or use a different software stack. I mean there is FHEM and some other projects.
And with that said, there is some merit to what you’re saying. Software should be designed with usage in mind. It’s just not easy and there are contradicting requirements. Either someone puts in all the effort to cater for your specific use-case… Or they don’t.
It’s not my method. Writing software with distributions in mind is the standard in free software development.
Indeed. That’s why many engineers don’t bother. Especially poor engineers.
yeah, you don’t listen.
LOL
You don’t have to use HomeAssistant if you hate it so much.