I remember way back when I was taught the difference between 'hardware' and 'software':
"If you can kick it, it's hardware, otherwise it's software."
... and that's the problem. Software, unlike some work of art lacks any physical form. You can't scrape off a landscape scene to reveal some hidden masterpiece - once the code is gone, it's gone for good.
Okay - yes - what am I doing, mixing 'art' and 'software' in the same sentence. Most everyone who looks at the majority of software knows it's nothing 'aesthetic' and certainly not 'art'. But that's not always the case. Kind of like those Microsoft ads, us geeks know that at some point in the future, software design will be seen as sexy. There will be a time when the creative process of solving such and such a problem, or designing that process will be viewed as something more refined than mere(!) engineering (or, possibly, considered on a par with traditional engineering). And by then, I wonder what will have been lost? How many designs, systems, applications, architectures, will have been erased, dumped, deleted, lost forever to the code heaven where all lost codes go? Already, how many pre-y2k systems were suitably elegant, and sophisticated to deserve more than the fate they suffered? True. Not all. Maybe not even 'many', but some - yes, certainly some I'd think.
So, should we be preserving these systems, rather than casually erasing them? How many 'legacies' should we be losing, and how many preserving? I don't mean 'maintaining', just moth-balling for future generations. For those yet to come. At least to learn from our mistakes (lest be destined to repeat them). I bet Microsoft still have Windows 3.0 available, but how many lines of business systems built in the 70s are still around somewhere? I know for certain that incredibly sophisticated and complex applications I have worked on are now no more. And literally no more. There presence lost from the software world; no longer in the physical world. No mark left, and no possibility of resurrection. Newer isn't always better; and even if it is better, then what came before isn't automatically irrelevant.
So, a museum of software. More than simply 'oldversion.com'. A celebration of what came before. A collection worth treasuring. Worth keeping. A testament to the endevour that came before. Surely, we should not be casually destroying this art?