I however agree that's is not going to happen both because of the nightmare of relicensing 3rd party code, and because an OSS Win7 release with community updates and security patches (no telemetry) would probably displace a good number of Windows 10 installs in short time.
If closed source software that talks to the Internet is an inherent privacy risk, let's maybe start by focusing on Google Chrome and Android, which have access to much bigger pots of honey than Windows itself does these days.
In short, this seems to be one of those spots where all the FSF has is a hammer.
If Microsoft was as dedicated to open source as their recent PR pivot and thrust has lead us to think then they would have no problem taking on this monumental task.
this 'stupid' PR move from the FSF is only a response to a similiar PR move from MS.
This just looks silly.
Blender also used to be a wholly closed-source product; it was opened up after the source code was "liberated" via a Kickstarter-like crowdfunding initiative - and it's now a successful FLOSS project. It can happen.
But just imagining that happening... Would be super crazy.
Instead, Microsoft has been virtually forced to continue to include the VB runtime DLL in every version of Windows since, because there are so many businesses out there that rely on software developed in VB6 that can't (easily or cheaply) switch to something else.
Windows 10, from what I understand, almost became the first version to ship without it, but Microsoft relented under the pressure (maybe I'm misinformed here, though).
VB.NET is sufficiently different enough that conversion either isn't possible, or due to the source code not being available (or the 3rd-party company that wrote it not existing any longer as an entity - or the source code assets not available or whatnot due to dissolution) - conversion becomes nearly impossible, short of a very expensive reverse-engineering effort by another 3rd party.
People's time isn't free.
The other problem is open sourcing the Windows code would be also involve open sourcing its crazy huge build system and tooling — it’s not a simple makefile — which is probably another level of impractical.
They were able to get rid of a lot of Windows 1.0/Dos back-compat junk with the release of x64 (due to loss of native 16 bit support). I'm sure there's still some though.
My employer uses 32bit VDIs to run a few legacy applications.
It must be pretty expensive to port/test all of their patches/updates on something few use [0.70% per [0]). They already killed it on their server line (although it was silly on there with the low memory limits). Nvidia has already announced they're ending support[1].
[0] https://www.pcbenchmarks.net/os-marketshare.html
[1] https://www.extremetech.com/computing/267180-nvidia-ends-sup...
I noted in another comment that they have a similar problem with people wanting VB6 open sourced, because so many businesses rely on VB6 apps - some written internally or exclusively for their business.
Windows 10 (from what I remember) almost didn't support VB6, but in the end they included the runtime DLL to allow for the apps to work. I wonder if this is a part of the 32x thing? Seems a likely possibility...
VB6 and the companies relying on it might be an albatross around Microsoft's neck; for how much longer, I'm not sure.
I find it difficult to believe that Microsoft would keep Windows 10 x86 around if usage was really that low.
Backblaze found that about 4.5% of their customers use 32bit Windows in 2017[0].
And this is if Microsoft DID want to open source it, which if they were giving in to a demand from FSF, or trying to do it as part of a goodwill "Microsoft <3 Linux" initiative then you will quickly find that the costs far outweight the benefits.
There are probably lots of other reasons that MS won't open source an older version of their flagship product, but this "OMG licensed 3rd party code" excuse is overplayed (by lots of other companies, too) and needs to die.