"Reliable Computing" and post-Redstone Windows

Avatar
4

I’m sure most of us are familiar with Paul’s repeated call for a “reliable computing initiative.” This has me thinking about overall stability and reliability in Windows 10. I like the way Windows 10 works, and it has gotten better over time, but I am annoyed by a lot of the little inconsistencies and problems that continue to crop up. However, I’m interested what a reliable computing effort might look like beyond taking a break from the update stream in favor of some consolidated bug fix patches (I think that’s kind of what Paul has in mind?)

Some of you may remember the release of Apple’s Snow Leopard back in 2006. That release was billed as one focused on refinement and stability as opposed to adding new features. Of course, there were some very significant new features and changes, but they mostly served the goal of improving the base system rather than adding new capabilities. A good example is rewriting the Finder in Cocoa – this might be compared (imperfectly) to rewriting File Explorer in UWP or .NET to gain overall system stability. There are a lot of things like that that MS could put together in a new version of Windows – polish, stability, and future-proofing sorts of improvements.

Anyway, I’m curious in whether people think such an effort is needed, whether it’s likely, and what form it should take. Stop rolling out new versions or feature patches and just work on bug fixes for some amount of time? Make the next release (or two) of Windows after Redstone 3 focused on those things a la Snow Leopard but continue new releases? And what are some of the most-desired changes of that sort? Personally I’d love to see a concerted effort on this front, but I’m assuming that Redstone 3 is already set in stone as another feature-oriented release.

Comments (4)

4 responses to “"Reliable Computing" and post-Redstone Windows”

  1. Avatar

    5664

    Considering it's possible for UWP apps to "break" and be left in an unworkable state, and that this can happen to the Store app... and the only way to fix it is to create a new user profile... then I would say it's definitely needed.

    There seems to be an organizational need for Windows to have some incredibly overly complex architecture, and nobody wants to clean it up. They really could do well by rearchitecting, and building a fully modern NT personality for UWP execution, rather than building it on top of the existing Win32 userland. Remember that NT used to run OS/2 binaries, had a UNIX subsystem once, and now has a Linux subsystem. Just build a first-rate UWP subsystem.

    • Avatar

      6852

      In reply to jimchamplin:

      Yeah. Not sure if this is related to the problem you describe, but my mother in law had a problem where the Photos app just doesn't work any more. You can click it but it will never launch. I spent hours googling, doing all sorts of arcane stuff... never fixed it. Consensus was "reinstall Windows 10."

      Shortly after that all of her Quick Access links just disappeared from File Explorer. Not meaning you can just put the links back, but meaning Quick Access is completely gone as a section. Why? Who knows. And this is all on a less than 1 year old machine - $1200, nice convertible laptop - that they use for nothing but surfing the net and storing/viewing photos from their camera. 

      So I'd love to see MS focus on just making sure apps and various pieces of the OS stop breaking at random. That'd be a good first step. And then of course all the other little polish issues and improvements they could make instead of worrying about stuff like further improvements to Windows Ink or whatever.

  2. Avatar

    5530

    I remember Microsoft actually had a re-written Explorer.exe completely in .NET but it was found to crash too much so they gave up on it.

Leave a Reply