There’s a bunch of costs to releasing closed source code as open source. Generally they’d have to have a team of developers review it, they’d need their lawyers to review it and any licenses. That every company that had licensed code used would also need to go through their own process.
At any time, one of them could say no and the process stops. So you’re basically paying for everyone’s time and possibly to relicense some parts. And im sure I’m missing other parts.
It's unfortunate that when a company DOES chose to release old software, the entire revision history is thrown out due to processes like that. In the software world, this probably doesn't matter too much, but video games can go through massive changes during development. But it also depends on factors like if game assets were even stored in version control in the first place.
Yep, it's a huge time sink even in the best case scenario, let alone with ancient software encumbered by transfer of ownership through multiple corporate custodians.
It's not that unusual for it to not even be wholly owned by a single entity due to licensing requirements.
It seems dumb and it is dumb, but it’s also an important feature of Windows that your terrible app from the 90s can still run today with minor changes generally. Less so if it’s win16 but anything win32 is good to go.
Also, Linux has a similar philosophy as Linus has eloquently described in the past. [1]
> It seems dumb and it is dumb, but it’s also an important feature of Windows that your terrible app from the 90s can still run today with minor changes generally.
This is a myth which is mostly not true from some time.
It's not like this kind of stuff can't be found by reverse engineering. For Windows specifically, the ReactOS team is doing lots of it to document the inner workings of internal Windows APIs so they could reimplement them.
> Is there a risk those closed source projects would be sued following the release because they disclosed some forgotten sin?
I know nothing about law, but I guess software companies would protect themselves in advance by mandating that all their developers use only software and sources that belongs to them anyway. That would be useful in case of a lawsuit where a court could force them to show the sources to prove they don't come say from the claimant's reverse engineered software, or some other FOSS project. Things might be different with 3rd party modules and libraries, where an obscure module was licensed from a company that went out of business ages ago, but all their assets were acquired by another party who finds the module in the wild and decides to sue.
Just speculating, but although I'm 100% in favor of FOSS, if the above were real possibilities, I couldn't but understand why some companies are so reluctant to release even old code they couldn't profit from anymore.
When developing software, a company might aquire a license to add some library or code to their product.
This happens all the time with OS's like Windows (licensing the ability to play media formats, etc.) or games (licensing certain engines or algorithms).
That doesn't mean the company squired a license to release the libraries or code as open source 20 years later.
So you'd have to go through the codebase with a fine-toothed comb, line-by-line, because the original developers don't work for you anymore, and the license agreements are long lost.
Releasing large software projects are a massive cost to any company that wants to be protected from potential lawsuits.
I'm thinking mostly of anti competitive practices from Windows.