Not again MS! Faster Insider Builds?
Apparently Microsoft is planning to ramp up the speed at which it releases Insider builds to fast ring to be closer to what their internal testers get.
Frankly, I think this is a bad idea. I want faster builds. But I still think that they need a new ring for them. And I think they need more confirmation prompts to get into it.
The first problem I see is that not everyone who is a part of the Windows Insider program is necessarily paying super close attention to such announcements. And some people are just plain idiots. So, people who aren't expecting it may be getting drastically different quality software than they are used too/expecting. Current fast ring builds tend to be pretty darn stable (and none too fast). Personally, I never needed to restore any of m PCs.
Next, by not adding another ring, it becomes a lot easier to rationalize going back on this decision. One catastrophically bad build and the decision might end up being undone.
As I've argued in the past, the way to do this which feels right is to add a 3rd ring for insiders. They joked about calling it the "Ludicrous Speed". I think it is a great name. It isn't wholly serious, but then people who want to seriously use the OS daily, probably shouldn't be using those builds.
As far as deterring people from actually subscribing to the faster build? Make your first warning: "This may result in daily OS updates that could take several hours to install". The thing most likely to make me opt out of such a ring would be daily updates. I suspect others feel the same. Your computer can be unusable for an hour or more each day. Add other warnings as well and most people will walk away.
Beyond that... I've also often advised demanding some sort of proof that you have a secondary computer that isn't in the ultra-scary ring that you could use to restore a machine if an update took it down. Not unreasonable measures. And these could address what I felt were the biggest problems with the existing Insider program. Primarily that holding off for something that met the Fast Ring requirements often meant not giving users enough time to provide feedback on new features or changes. Which, is supposedly one of the core purposes of the Insider program.
So, yeah! Once again, I think they goofed. I suspect this proposed fast ring change will be short lived. We'll get a couple of rapid fire builds. Eventually a serious bug will slip past, hose a ton of machines and they'll revert back to the old process. Again.
The other reason I disagree with doing this in the Fast Ring is that I think that there is a well defined purpose for 3 separate rings:
Here, Ludicrous speed allows for VERY rough features and functionality to get out. This would be for people who truly want to live on the edge.
Microsoft gets the most out of this ring. Testers who don't mind restoring machines to test the install/update process in real world scenarios. More hardware configurations and setups to catch bugs on and super early feedback for features and changes.
Users get an interesting new insight into the development process. Seeing features come and go and contributing in a very meaningful way to the product.
Fast ring would be for people who don't mind installing a new build every week or two and don't mind a few bugs but still want some reassurance that there is a good chance that they won't need to restore their machine. Features which make it into the Fast Ring builds should be considered very likely to remain in the final product but perhaps altered based on feedback.
Microsoft gets an expanded ring of testers to find less common scenarios which may not have been uncovered in the smaller ludicrous test ring. Also, serves as a potential restore point for Ludicrous Ring testers to revert to if install breaks their build.
Slow ring would be like a Release Candidate level of software. From a user perspective, you want things a bit earlier, but you don't want super frequent updates and don't want unstable code. You can tolerate rare/known bugs if you have to, but only in extreme conditions should a restore back to a GA release be required.
For Microsoft, this just expands the test group even further to find more fringe cases. Also, a restore point if a Fast Ring build breaks someone's system. Should be highly unlikely that features here don't go into GA release. By the same token, this late in the process, feedback is less likely to get incorporated in the current dev release.
Frankly, I think this is a bad idea. I want faster builds. But I still think that they need a new ring for them. And I think they need more confirmation prompts to get into it.
The first problem I see is that not everyone who is a part of the Windows Insider program is necessarily paying super close attention to such announcements. And some people are just plain idiots. So, people who aren't expecting it may be getting drastically different quality software than they are used too/expecting. Current fast ring builds tend to be pretty darn stable (and none too fast). Personally, I never needed to restore any of m PCs.
Next, by not adding another ring, it becomes a lot easier to rationalize going back on this decision. One catastrophically bad build and the decision might end up being undone.
As I've argued in the past, the way to do this which feels right is to add a 3rd ring for insiders. They joked about calling it the "Ludicrous Speed". I think it is a great name. It isn't wholly serious, but then people who want to seriously use the OS daily, probably shouldn't be using those builds.
As far as deterring people from actually subscribing to the faster build? Make your first warning: "This may result in daily OS updates that could take several hours to install". The thing most likely to make me opt out of such a ring would be daily updates. I suspect others feel the same. Your computer can be unusable for an hour or more each day. Add other warnings as well and most people will walk away.
Beyond that... I've also often advised demanding some sort of proof that you have a secondary computer that isn't in the ultra-scary ring that you could use to restore a machine if an update took it down. Not unreasonable measures. And these could address what I felt were the biggest problems with the existing Insider program. Primarily that holding off for something that met the Fast Ring requirements often meant not giving users enough time to provide feedback on new features or changes. Which, is supposedly one of the core purposes of the Insider program.
So, yeah! Once again, I think they goofed. I suspect this proposed fast ring change will be short lived. We'll get a couple of rapid fire builds. Eventually a serious bug will slip past, hose a ton of machines and they'll revert back to the old process. Again.
The other reason I disagree with doing this in the Fast Ring is that I think that there is a well defined purpose for 3 separate rings:
Here, Ludicrous speed allows for VERY rough features and functionality to get out. This would be for people who truly want to live on the edge.
Microsoft gets the most out of this ring. Testers who don't mind restoring machines to test the install/update process in real world scenarios. More hardware configurations and setups to catch bugs on and super early feedback for features and changes.
Users get an interesting new insight into the development process. Seeing features come and go and contributing in a very meaningful way to the product.
Fast ring would be for people who don't mind installing a new build every week or two and don't mind a few bugs but still want some reassurance that there is a good chance that they won't need to restore their machine. Features which make it into the Fast Ring builds should be considered very likely to remain in the final product but perhaps altered based on feedback.
Microsoft gets an expanded ring of testers to find less common scenarios which may not have been uncovered in the smaller ludicrous test ring. Also, serves as a potential restore point for Ludicrous Ring testers to revert to if install breaks their build.
Slow ring would be like a Release Candidate level of software. From a user perspective, you want things a bit earlier, but you don't want super frequent updates and don't want unstable code. You can tolerate rare/known bugs if you have to, but only in extreme conditions should a restore back to a GA release be required.
For Microsoft, this just expands the test group even further to find more fringe cases. Also, a restore point if a Fast Ring build breaks someone's system. Should be highly unlikely that features here don't go into GA release. By the same token, this late in the process, feedback is less likely to get incorporated in the current dev release.
Comments
Post a Comment