Then what's your definition based on your own goal posts?
Making software that's sitting around and doesn't generate any revenue? That's either a hobby which is fine, or is a waste of resources, both in development and in storage/deployment.
And if the 20 year old software isn't unmaintained but is actually constantly getting updated and support, then is it still 20 year old software or is it basically up-to-date software (the Ship of Theseus problem)?
It doesn't have to produce revenue, it just has to aid in the production of GDP. I don't pay for Linux, but it is a key part of the GDP I am responsible for.
And while Linux is maintained, the vast bulk of the value it provides is quite old and almost unchanged.
But the Linux we use today isn't 20 year old software. Sure, the foundation is over 30 year old, but most of the Linux code we use today is relatively fresh and is constantly being improved and updated. Again, I bring the ship of Theseus argument. I doubt my Android phone has too many lines of code written by Linus on it.
Not many places are running software that saw no more active development for over 20 years. Even governments and banks are pushing updates to ancient codebases.
Unmaintained 20 year old software would be the firmware running on your 20 year old car or microwave oven, as that's frozen on that system, not getting any more development.
So, out of all products made by humans, software has by far the shortest expiration date, with the exception being food products.
You completely misunderstood my point. Yes bit-rot exists, but that's not what I was talking about. Software doesn't rot like food, it just stops being useful for its intended purpose if left un-updated for too long.
Just look at my 10 year old Android phone. It still boots and works just fine, but having not received anymore SW updates for 8 years, it's completely unusable for its intended purpose today despite it still running. There's nothing wrong with the hardware, OLED screen is still very bright and sharp and I changed the battery and everything, but with such old unmaintained software it's basically an e-waste paperweight now.
Most software doesn't stay useful forever, and without updates it has a shelf-life, as most software becomes useless after a while if left un-updated. So yes, my point stands, old un-updated software provides no value in most cases.
Making software that's sitting around and doesn't generate any revenue? That's either a hobby which is fine, or is a waste of resources, both in development and in storage/deployment.
And if the 20 year old software isn't unmaintained but is actually constantly getting updated and support, then is it still 20 year old software or is it basically up-to-date software (the Ship of Theseus problem)?