I've literally seen a Rust shill promoting the idea of adding dependent types to Rust. Spoiler: You do not want dependent types in a language that's (ostensibly) used for writing practical programs, since you have to either restrict the language to not be Turing-complete to make type checking decidable, or else accept having a type system where type checking may never terminate.
considering that things like c++ template expansion may never terminate, i wouldn't say dependent types are a dead end off the bat like that
I mean, we are talking about someone willingly choosing to write a web forum in Rust so it's kind of on-topic!
That's literally what the Programming Thread is for.
there is a programming thread, but there isn't a programming language thread
it's for programming, which technically includes programming language theory but maybe not?
People will complain its not on topic and the resident lolcows/autistics won't use a containment thread in mass debates.
in the future hopefully we will be able to use hardware produced from free designs that can run with free microcode (
)
Exactly. Don't get me wrong, I love the FSF and what they stand for, I just think their approach is a little bit too rigid, at least for my use case. Their primary concern is software freedom, but where security is concerned, being pragmatic overrules the adherence to 100% Libre everything. Kernels & Libreboot are good examples. Yes, both allow for some proprietary microcode to run in the system, but their implementation is beneficial enough to look past those things. At the end of the day, Libreboot lets you effectively nullify the IME spyware issue and LTS/mainline Kernels let you mitigate potential vunerabilties, and that to me is worth the presence of whatever blobs could not be replaced or removed.
There are FPGA FOSS projects, but they're relatively recent. If you're interested in autistic, schizophrenic setups Chipsalliance has been funding FOSSi FPGA project. I can only tell you this much without a self-dox.
i actually like the extreme rigidness somewhat. extremely strong principles are fucking based. "this repository has NO proprietary software at all and if you don't like it maybe you should install a repository that does have proprietary software"
some people might be running an airgapped system with a threat model that allows not running updated microcode. trying to railroad them into using proprietary software when they don't need to use wouldn't be great
whatever relation you have to them, good luck
That is why I love Guix's approach so much. The base install adheres to those principles, but modifying it to fit your specific needs, free or otherwise, is extremely easy and intuitive. People should never be forced or railroaded into running anything, I was just giving some commentary on when and why I think some microcode is better than purist 100% libre-ness. But, like you said, if I could, I would run it without one speck of propriejeet shit.
They have to be rigid. If the free software foundation is make exceptions for some piece of proprietary software. Then that sets the precedent to do it again in the future. And if the FSF is promoting the use of proprietary software, where does that leave the standard that is being set?
observant readers will notice that their name is this, and not the "free software and a few bits of proprietary software here and there where it's necessary to make this hardware work and be maximally secure foundation"
Perhaps I should have been more careful with my wording: the FSF's guidelines are too strict for my specific use case. I absolutely agree they should remain purist. Were it not for their rigidity FOSS would probably be in a far worse state than it is today, and I ardently support them wholesale, even if I have to sometimes go against their grain.
