“GOOD EVENING, ID LIKE TO HAVE AN ARGUMENT PLEASE” lol
Not much I would say wouldn’t cover ground others have covered. I wouldn’t have thought wine/dxvk/lutris would have come as far as they have and it’s truly impressive.
One upon a time there was a spreadsheet program called Lotus 123. It had 99.9% of the spreadsheet market. Microsoft put 10x the developers working round the clock to unseat it and get their foot in the door in businesses with Excel. From Almost the very beginning Excel could open 123 formats.
It wasn’t until Excel 3.0 that Excel could save back to the 123 format. Almost overnight Excel supplanted 123, which was an unexpected result, and meant no one needed the ‘save as 123’ function. And yet it was necessary for the transition from 123 to Excel to be possible. It took < 1 year after this change for Excel usage to surpass 123. In an age before the internet, that’s crazy fast.
This result was so baffling at the time (now obvious) it’s the reason Microsoft did all the FUD stuff with office formats in the 2000s. The office formats are so complex it’s an impossible hill to climb. Even now, 20 years later, open programs struggle with Ms formats (even with open documentation! ) for any complex document.
This study of office formats and how all that went down will play out like replacing windows over the long term, I suspect.
Programming to replace windows is a godawful tarbaby. What if I told you instead of getting in there and getting your hands dirty, doing companies jobs for them attempting to black box port their code to another platform, you could just wrap up the tarbaby in a sandbox and put the ball back in their court to solve? A cync might say that it’s in Microsoft’s best interests for the best and brightest developers to be busy with wine/dxvk/lutris (if they think their overall platform is an equally landmine-laden impossible hill to climb as is their file formats) rather than focusing those developers on shoring up other weaknesses on Linux, as a platform, “native” functionality.
I don’t have to disagree with some of the anti-vfio points but nothing will persuade me that vfio enabling users to ‘save as 123’ will lead to the flashpoint of Linux on the desktop.
With tech like looking Glass leading to not just desktop forwarding but individual app forwarding – this is the future. Especially when the tarbabay gets aggro’d. (Have you even thought about a future where Microsoft becomes Oracle wrt the caselaw precedent Oracle is setting on APIs? Have you noticed how much “Install Visual C++ Runtime from microsoft” WINE requires?)
It will take millions of hours to perfectly reproduce windows insanity in wine/lutris/dxvk and doing that with out software vendor cooperation would be… tricky. We do have valve leaning on their publishers but vfio has already enabled legions of folks to ‘no compromises’ switch to linux on the desktop 100% of the time.
This happened once before – windows terminal services. As computers got more powerful, it became practical to cram a million users onto one box. Cheap thin clients that didn’t need windows popped up. Linux was actually making a lot of headway then – run what you can locally then run what you can’t via terminal services. The best multiplatform of both worlds. Guess what MS did then? “Clarified” that they require a client license on the device connecting to terminal services, in addition to the terminal services license. In business settings, for terminal services. The cost of the software on that client? Same as a windows license, which was free if your client software was windows. Ouch. Shady.
So, long term, I worry that Microsoft still has some sort of legal capability to pull the rug out after a ton of development has been invested. Whereas, legally, any kind of license maneuvering to prevent you from running an unmodified licensed copy of windows in a VM would probably be dimly viewed from an anticompetitive standpoint the reverse – “the heathens stealing our proprietary APIs” (if that were to be the argument from MS) would probably afford ?Microsoft much more legal protection. (Look how far Oracle has got. Scary times right now. )
Either way, interesting times are ahead. Unless Microsoft willingly gives up the desktop OS in favor of massive cloud services profits (which could very well happen) I don’t see the “Year of the Linux Desktop” happening without vfio. (Though I’d welcome it with open arms if it did.)