Open Source software that enjoys continued development by the broader community obviously works differently; this is about proprietary software that will not be ported to new platforms, receives no security fixes over time, and will just stop evolving at a given point in the past.
TeX would not be as useful if the only version you had available would be one Knuth compiled himself in 1990, without the original source code; and the fact that TeX solves a contained problem,
he intended it to be a „finished“ project, and was an exceptional developer makes this a good example to prove your point, but a bad one to counter mine: most software is just not like that.
To the contrary; a lot of software is highly contemporary, depends on other, external things, and requires a constant stream of security updates to cope with the rest of the world. This causes software to loose value rather quickly.
> this is about proprietary software that will not be ported to new platforms
Software being proprietary is entirely a choice of the vendor. There's nothing inherent in MS software that means it has to be proprietary. It's a pretty essential part of planned obsolescence.
As for new platforms, what new platforms? I can still buy a complete system that can execute x86 code that was written decades ago. And, in any case, we've been able to write platform agnostic code since before many programmers were born.
Not all software needs to care about the "rest of the world". There is a ton of software that doesn't need any kind of network support at all to deliver value. The only real reason for software to become less valuable is if the problem it solves goes away. This can happen, of course, but I don't think it happens as much as MS and other want us to believe.
Software that is designed well is able to be adapted and re-used for new purposes. TeX didn't have good support for separating style from content. But we didn't say "TeX isn't valuable any more", people simply built further tools on top of it like LaTeX and ConTeX etc. This is also why Unix remains just as valuable as it always has been. Unix surely delivers more value to more people today than it ever has done. It turns out the world really doesn't change as much as we sometimes think it does.
TeX would not be as useful if the only version you had available would be one Knuth compiled himself in 1990, without the original source code; and the fact that TeX solves a contained problem, he intended it to be a „finished“ project, and was an exceptional developer makes this a good example to prove your point, but a bad one to counter mine: most software is just not like that.
To the contrary; a lot of software is highly contemporary, depends on other, external things, and requires a constant stream of security updates to cope with the rest of the world. This causes software to loose value rather quickly.