3 ms·

They also have very large software footprints, and enough bugs in their latest/supported versions that it is difficult to justify taking resources from latest t

They also have very large software footprints, and enough bugs in their latest/supported versions that it is difficult to justify taking resources from latest to support to older versions with dwindling user base. I know the software upgrade cycle is frustrating but if you’ve ever tried to both release new features and maintain legacy versions of software you know how complex that quickly becomes. It is not simply a matter of releasing legacy source code. Modern software is extraordinarily complex and interwoven. Modern software doesn’t just run in one place, it exists across a multitude of systems and services, each with their own contracts and interfaces that evolved over time. There is a reason these companies have tens of thousands of engineers maintaining their code bases. Not to mention there are very real concerns about proprietary information, as many components of legacy systems are present in derived forms in their successors. I’m all about right repair - especially for hardware - but it’s a bit naive to think that we’ll find nirvana in opening up legacy source code. I know it’s a tinkerer’s dream to dive through old versions of software you used to run and add your own little personal features, but realistically it would do nothing to move the needle much on actual public use of legacy software. Don’t forget the average user often cannot even troubleshoot why their printer is not working, let alone install, run, and troubleshoot a custom android stack on their 2006 Nokia.

agentultra 2y ago
I've been a professional software developer for over twenty years and writing software since I was a kid in the late 80s. I understand why it's hard supporting old platforms. The source code is worthless. It's just one way a company could extend support for older platforms they've given up on supporting. It's not valuable for them to hold on to the sources. We could all huddle around the existing Unices and BSDs and try to keep the forks going for older platforms. That's fine too. But the original sources may have hints about how to support some of the custom hardware for a given platform that would be useful, etc. It really wouldn't be hard for them to just say, "Okay, here's the source for the last build 12.x.x known to run on hardware revision 1.x.x," and leave it at that. Doesn't cost anyone anything. Update: What's trickier perhaps is supporting older platforms through emulation and only having binaries. Just open the source, Luke.

agentultra 2y ago

I've been a professional software developer for over twenty years and writing software since I was a kid in the late 80s. I understand why it's hard supporting old platforms. The source code is worthless. It's just one way a company could extend support for older platforms they've given up on supporting. It's not valuable for them to hold on to the sources. We could all huddle around the existing Unices and BSDs and try to keep the forks going for older platforms. That's fine too. But the original sources may have hints about how to support some of the custom hardware for a given platform that would be useful, etc. It really wouldn't be hard for them to just say, "Okay, here's the source for the last build 12.x.x known to run on hardware revision 1.x.x," and leave it at that. Doesn't cost anyone anything. Update: What's trickier perhaps is supporting older platforms through emulation and only having binaries. Just open the source, Luke.