Reply to post: Re: Seduced and abandoned by yet another attempt at a non-x86 Windows?

Re: Seduced and abandoned by yet another attempt at a non-x86 Windows?

For me, Google is still the leader when it comes to offering stuff to developers and then leaving them high and dry later. Apple also has history of abandoning developers at dead-ends with no migration path except “rewrite everything” (OSX CarbonLib?), although it has improved immeasurably if you compare now to its Don Juan attitude in the 1990s, when APIs would be trumpeted as the Second Coming at one year’s WWDC, then completely written out of history by the time the next one arrived (Just off the top of my head I can remember OpenDoc, Dylan, MacApp, OS 8 “Copland”, Appearance Manager, QuickDraw GX... and I may even have contributed my own tiny part toward this too while at Apple)

Returning to Silverlight as the example, at least there had a forward migration path to converting apps into WinRT 8.1 applications - a class-path rename would get you about 75% straight carryover of application code, and Microsoft went to some effort to provide you with workarounds when that didn’t work, which isn’t bad considering to complete platform change underneath. Sure, you lost the opportunity to run in-browser, but by that time using these browser-plugin runtimes was rightly seen as far less efficient, and no more secure, than just downloading and running a signed application.

The issue for users with these changes is that Microsoft’s third-party devs are pragmatic when it comes to the use of their time, and are slow to rewrite apps just so they can say they are using The Latest Library (in contrast to Apple’s cohort who tend to leap over themselves to be first to have rebuilt their app against whatever Apple announces).

That’s not saying that Microsoft’s current clusterfuck with GUI toolkit APIs is an any way okay, mind you. I happen to really like WinRT, or UWP as it’s known these days, (and I especially like its responsiveness on slow CPUs), but I’d be happy to abandon it and use something else that has a long-term future if only MS would just make its damned mind up about what that “something else” should be.

As for x86, Microsoft isn’t to blame for these failures. Intel has very effectively used its position as the dominant supplier of desktop and laptop chips to discourage vendors from producing non-x86 Windows machines. Remember when Windows NT was launched on about five ISAs: Microsoft really wanted to break with “WinTel”, but the hardware builders weren’t so free to do so. Same story today - x86 ISA chips have higher price and higher profit-per-unit than ARM parts, and so the suppliers of these really want hardware vendors to stick with x86. Maybe if AMD applied some of its Ryzen know-how to an ARM ISA part, we’d see real competition, but I wouldn’t hold my breath.