However, at my previous job I integrated an LFBGS[1] implementation into our production code.
That was written in the early 80s in Fortran '77.
The code outlived all hardware that it ran on when it was first written (we ran it on an OpenMP cluster for a scientific-computing problem related to lithographic mask optimization), if not its authors.
It will continue to exist, and run, for a very long time.
Sure, all software is ephemeral. But as you say so, you probably used SciPy/NumPy. Deep inside, there's an implementation of LAPACK/BLAS doing the heavy lifting for you[2]. It started in 1979, and is still kicking.
Wow, I'm actually using some of the C++ code based off of MINPACK for one of our current projects at work. It is quite humbling to realise this dates back so far.
Unless you're a game developer! Lots of people are still playing Super Mario Brothers daily! They're running it in an emulator now, most probably, but all of the original code is doing the work.
Obviously it won't outlast the heat death of the universe or anything, but I'd say it's has the potential to live as long as any other human creation.
That's a grand assumption :) Who knows whos GitHub code will be included in the "Big MasterScript" that is written into the consciousness of the universe in the future, making it forever.
I know, far out there and in theory, I agree with you, everything is temporary and nothing is forever. But who knows, maybe in the future, things will no longer be ephemeral.
I think that thinking this way is a coping mechanism in response to the massive loss of control we experience being subject to the whims of giant corporate platforms deciding what we can and cannot run on our own devices.
The alternative is to face how insanely unfair and belittling it is to have bought an app and a device and due to circumstances out of your control arbitrarily no longer be permitted to run “your” app on “your” device.
You think that being aware of our own inevitable deaths, and the eventual end of the universe, is a coping mechanism to avoid the unbearable reality of closed software platforms? Am I talking to Richard Stallman?
If humans discover a way to defeat entropy, even at a limited scale, whatever device is created to preserve information into the era of a universe solely made of black holes and hawking radiation will be capable of running doom.
I fully expect the last computer to ever be built by humans to still be able to load original .wads.
However, at my previous job I integrated an LFBGS[1] implementation into our production code.
That was written in the early 80s in Fortran '77.
The code outlived all hardware that it ran on when it was first written (we ran it on an OpenMP cluster for a scientific-computing problem related to lithographic mask optimization), if not its authors.
It will continue to exist, and run, for a very long time.
Sure, all software is ephemeral. But as you say so, you probably used SciPy/NumPy. Deep inside, there's an implementation of LAPACK/BLAS doing the heavy lifting for you[2]. It started in 1979, and is still kicking.
(And it's still in FORTRAN)
[1]https://en.wikipedia.org/wiki/Limited-memory_BFGS
[2]https://en.wikipedia.org/wiki/Basic_Linear_Algebra_Subprogra...