5 ms·
A reminder to invest in libre/free/open infrastructure folks. It's not just widening the tent and lowering the barrier to entry for people today, it's also maki
by abetusk 5y ago
A reminder to invest in libre/free/open infrastructure folks. It's not just widening the tent and lowering the barrier to entry for people today, it's also making it accessible to all of our future selves.
- wiz21c 5y agoWell, free software also do die sometimes. But sure, it's by far the best option. All the things I do with a computer is based on free or open source software or open standard. Curiously, everything still works 20 years after... Watching movies, listening music, text/calc, coding, mail, internet,... (not playing game though, the kids have their console but they don't create anything with it, so I don't care if it gets obsoleted)
- praptak 5y agoFree software dies but usually because of lack of interest. Case in point: Ruffle. Flash died, free software picked up the torch.
- ant6n 5y agoFree software won't mean that operating systems won't disable software by retiring APIs. For example on Linux, at some point somebody deprecated some systray icon Api. And some linuxes, like elementary OS just full on disabled systray icons, with some arguments about this having been deprecated years ago, and u shouldn't use software that depends on systray icons. This doesn't just disable old software, but also current software - for example the Dropbox client doesn't work out of the box, and caffeine/caffeinate doesn't work either.
- abetusk 5y agoYou bring up valid points about engineering practices and backwards comparability in general. I do think you're nitpicking a bit when it comes to the issue at hand. For example, in you grievance, is there a way to provide a Docker solution to run the working code under 'emulation'? Is there a way to compile an older version of the kernel to run these programs in the way you want? I don't know but my bet is that the answer is 'yes' to both of those. My point is that this wouldn't even really be possible without a FOSS ecosystem. It requires labor from someone who cares enough to fix it but at least the avenue is open enough in the form of accessible standards and code. Under a proprietary standard or source these would need to be reverse engineered or, in maybe an extreme case, be illegal.
- nikanj 5y agoFor a professional programmer, OSS from 20 years ago is not a hopeless task. Read compiler errors, fix headers, patch things, eventually get a working build. For a non-technical person who just wants to get their animations working again? Essentially impossible.
- password4321 5y agoQt4 is already too tough for me!
- abetusk 5y agoThis is a valid point but there are a few counterpoints that come to mind: * FOSS now, while complex and intimidating, at least has the possibility of allowing for bug fixing and alteration (and having those alterations fold back into the main project) whereas proprietary software almost surely does not. * The situation is not as dire as you suggest for many projects, especially ones that are self contained or smaller in scope. Arduino libraries can be understood and updated. There's a lot of scientific software that can have parts isolated, updated and fixed for someone knowledgeable enough. Technical tasks like fixing headers, patching code or even using Git, are intimidating and difficult. Many projects are so complex now that on-ramps, even for highly technical users, are difficult. I don't see that changing anytime soon. FOSS at least offers people who can understand these technical tasks the possibility to take actionable response. Convenience is absolutely something that should be prioritized but proprietary software and closed standards favor ease of use at the cost of freedom and, ultimately, longevity. My view is that FOSS is a necessary but insufficient element to a healthy ecosystem. FOSS, especially compared with proprietary software, is at least a step in the right direction.