Future of Windows 10 on arm

[clintroymkt]

I often see the true long term benefit of windows 10 on arm being that in say a few years , we can have windows versions of really cheap android phones, maybe not 50 dollars but perhaps something akin to a core i3 machine for around less than 200 dollars. it would be a huge help for people in countries like mine (Zimbabwe) and could help microsoft be part of that push for the next billion internet users.

 

[Brett Watters]

First, you owe us a definition for "good enough". I'm going to assume you define "good enough" as:...

I have no idea what that point is. It obviously varies. I still use a 950xl, so good enough for me is meaningless for most people. However, if you have an Andromeda device which can replace your tablet... and it can make phone calls, send-receive text messages, maps, GPS, etc. there is obviously a benefit in not also carrying a phone around if you can avoid it - weight, cost, hassle of two devices, etc.

Second, I don't know what your current understanding of UWP is, so I'm not sure if I can give you a succinct answer.

Computer engineer. I've written UWPs.

Yes, Andromeda will be able to run all apps in the Microsoft Store, that includes UWP apps, PWAs and Win32 apps that are distributed through the Store using the Win32 bridge, a.k.a. Centennial.

Proof? I find it highly unlikely it would support desktop apps - ignoring the extra code, power, performance, support, etc. issues - desktop apps are going to be unusable on a 6" device -- even 8" tablets a too small to use touch even if the application scales on a high-resolution screen.

Furthermore, whether the software library at your disposal is 20x larger than it was back in 2015 depends on how you count.

I'd say most UWP could run on small screens... the WP10 issues was that developers simply didn't have a WP or weren't going to spend the time to test or develop against. When Windows 10 allowed UWP to run in windows... most fit without half or quarter screens without issues. My 8" tablet runs 1200x800 and I've yet to see a UWP which won't run under it. A 6" device at 1080p (or equiv with scaling) is certainly not too small for these.

I would say that if it can replace a small tablet, or was otherwise a "must have" for pen, notes, art, etc. people using it will immediately want to drop their phone. What percentage can, I don't know. I'd (presumably) would have no issue since I don't (generally) care about apps and still carry a 950xl. With the full UWP library, full Edge, UWAs, etc I could see a lot more people than me doing so.

 

[anon(50597)]

I'd (presumably) would have no issue since I don't (generally) care about apps and still carry a 950xl. With the full UWP library, full Edge, UWAs, etc I could see a lot more people than me doing so.

I think this is a fair point and probably the demographic it will initially migrate to. It will be very niche, just like the amount of people using a 950xl. Will that be enough? Will it catch on with others? We will see.

 

[fatclue_98]

Does anybody here really think Windows will be running current Win32 programs 10, 15, 20 years from now? Ignition points and carburetors gave way to electronic ignition and fuel injections, respectively, as I suspect WoA will be the near-future replacement for what we have now. It took 75 years for points and condensers to go the way of the Dodo. God help us if it takes the same amount of time to get out from under Win32.

 

[a5cent]

Does anybody here really think Windows will be running current Win32 programs 10, 15, 20 years from now?

Casual IT users? Nope. In some corporate settings? Unfortunately... Yes... (most definitely if we take the 10 year mark).

I have no idea what that point is.

If you can't define what you mean by "good enough", then you logically can't ask whether something will meet that standard. Without a definition such questions are just meaningless word salad.

I'd say most UWP could run on small screens... the WP10 issues was that developers simply didn't have a WP or weren't going to spend the time to test or develop against. When Windows 10 allowed UWP to run in windows... most fit without half or quarter screens without issues. My 8" tablet runs 1200x800 and I've yet to see a UWP which won't run under it. A 6" device at 1080p (or equiv with scaling) is certainly not too small for these.

1)
Why do you say "most UWP software could run on small screens"? What does that even mean? Did you mean to say "can"? Do you mean that UWP software technically could but won't. Do you mean to say that UWP software currently doesn't but will in the future?

Please be more precise.

2)
The problem was not that developers didn't have a WP. $500 for a test device is problematic? C'mon!

The problem was that some companies were simply hostile towards MS mobile ambitions, and that for most independent developers, the sales and revenue potential, in particular for apps that can't achieve a global following, was too low due to low market share.

3)
Note that most of the apps in the store for W10M are WP8 apps and will not run on a modern UWP platform... none of those count towards the number of apps that could run on the device when folded (at least I've never heard anyone at MS claim otherwise).

4)
Note also that developer engagement with the Store is generally low, and the bit that MS does enjoy is almost entirely focused on bringing over Win32 apps with the help of Centennial. None of that software will run on an Andromeda device when folded either.

5)
I just don't see how you've come to your conclusion. Can you link to a few examples of non-MS UWP apps in the Store, in the strictest sense of the word (since you're a fellow professional software developer I'll assume you understand what I'm asking), that re-composite their UIs for small screens? A small screen is not a half or quarter of your desktop. A small screen is about 16:9, portrait, with roughly a 5" diagonal.

As far as I can tell there are very few UWP apps that can do so. If you're saying that may be true currently but will change in the future, then we're at an impasse because that's a prediction neither of us can prove or disprove. IMHO, until hundreds of millions own such a device, most developers will simply ignore it, just as they did W10M. I see no reason why developers should behave differently this time around, in particular since MS has already stated that this will be a niche device (at least initially).

Proof? I find it highly unlikely it would support desktop apps

In this debate, my position is that for the average smartphone user, the continued existence of the mobile app gap won't allow MS' foldable device to serve as a smartphone replacement. You postulate the opposite.

In terms of being a decent smartphone replacement, Win32 doesn't really make much of difference either way. If we were to say it did make a bit of a difference, and we assume you're right that MS' foldable device will not run Win32 software, then that makes the app gap larger, which supports my position. I'd rather see you arguing for your own position rather than mine. ;-)

I can provide you with a few links on Win32, but that's likely to lead to a debate as well, which for the topic at hand would be a distraction. We can return to that later.

 

[Brett Watters]

If you can't define what you mean by "good enough", then you logically can't ask whether something will meet that standard. Without a definition such questions are just meaningless word salad.

WP10 is dead. Can we agree that the app gap is the main reason for this? If so, do you think it would have died if it could run all UWP apps? If WP10 could run all UWP, it would have something like 5x maybe 10x the number of apps available.

Why do you say "most UWP software could run on small screens"? What does that even mean?

I even gave you an example. I have an 8" device. It can run every UWP I've tossed at it. Screen size isn't an issue and its running 1280x800. There is no reason standard UWP couldn't run on a phone, other than the fact that most developers never compiled it for mobile.

The problem was not that developers didn't have a WP. $500 for a test device is problematic? C'mon!

If only software were just buying the device. You need multiple devices because you need to test on different performance, screen sizes, hardware, etc. and have multiple developers/testers working on them. Next, some customers are going to have an issues with running it on something you don't have. Partner programs, betas, dealing with manufactures, etc. Then, you need to setup emulators for each of your developers. Then, you need to code the application to avoid any API with issues and any hardware issue on any device. APIs have to be checked and if not available coded around. Next, you need to maintain that code across every other code base. Every time someone makes a change to the code, adds a feature, fixes a bug... you have to run/test it on emulators and/or the actual hardware. Then of course, you need to keep developers familiar with UWP on mobile around... often for years. All that time and effort can be spent on other things... i.e. web, iOS, and Android apps.

The problem was that some companies were simply hostile towards MS mobile ambitions, and that for most independent developers, the sales and revenue potential, in particular for apps that can't achieve a global following, was too low due to low market share.

I don't think most companies cared one way or the other. Anyone writing a UWP is probably not anti-Microsoft. If it was as easy as hitting a switch and recompiling to support mobile... there would be no business reason not to do so. Yes, it is physically that easy, but of course, it isn't in practice. You have a mobile version and you get hundreds of support calls for obscure bugs, issues, etc. Mobile apps go out (cheap or free) to hundreds of thousands of users and is typically seen as a cost, not a revenue stream. Writing a UWP was already an issue -- Win32 runs on all PCs, UWPs don't run on Win7, so that's already an unnecessary cost.

Company I was with never compiled their UWP app for mobile. It was a pure business decision and it certainly wasn't over $500.

4)

Again, I need to see proof that Andromeda is going to support any Win32 apps. Doesn't seem likely in my opinion. Andromeda is built on Core OS, not Windows on ARM. Seems like a lot of work, code, etc. for little benefit. I also don't think Win32 apps would be usable even unfolded. Win32 apps on my 8" tablet is virtually unusable. Scaling on high-res screens would be even worse. Maybe for a continuum mode.

I just don't see how you've come to your conclusion.

What conclusion do you think I'm making?

Again, if there is an Andromeda device which fits in your pocket, does light computing, pen, inking, drawing, notes, etc. and it can make/receive phone calls... then it would seem obvious that you wouldn't want to carry a phone with you. Same way I don't carry my tablet when carrying my 2-in-1. Now... what percentage of people are willing to give up their Android/iOS apps for this... probably low. However, I don't think it is non-zero. I haven't made any conclusions other than to suspect that people won't want to carry this along side a phone and there would be a percentage of people (given enough apps) who would ditch their phone as it would be redundant.

Can you link to a few examples of non-MS UWP apps in the Store, in the strictest sense of the word (since you're a fellow professional software developer I'll assume you understand what I'm asking), that re-composite their UIs for small screens?

Flipboard, Spotify, Lastpass, any Edge extensions, simple Solitaire, Slack, etc. These were just ones which I have. None where compiled for mobile, yet all scale and adjust fine to smaller screens, portrait, or high-res screens (8" tablet or resized to the left half of a 4k screen). We also have sub 7" Windows 10 devices and UWPs run on these fine. And we have hacks of 950s running UWPs on Windows on ARM. I don't see why a UWP would have any issues on such devices. It's pretty hard to write a UWP without already considering that fact that the user can resize the Window or at least dock it to one side. If your app runs on a 4K screen docked to the left with say 200% scaling, you are pretty much perfect for most smart phones in portrait.

Again, I'm still waiting for proof that Andromeda can run unmodified UWPs. It's not running Windows on ARM, but Windows Core with CShell.

In this debate, my position is that for the average smartphone user, the continued existence of the mobile app gap won't allow MS' foldable device to serve as a smartphone replacement.

I agree.

You postulate the opposite.

I never said or even hinted at that.

Again, I've said that if this device is useful as a tablet replacement that there will be an obvious want to not carry a phone with you. It can make phone calls, text, emails, maps, calendar, etc. Cost, size, transferring information between devices, learning, etc. would all make compelling reasons to ditch your phone. The fact that you have this device probably means you aren't the average smartphone user. Drawing, inking, running Windows apps, etc. probably means you values these higher than the average phone user. You have some need which your phone already can't handle.

Given this, my only point is that being able to run all UWPs would be a major advantage over mobile apps or apps which rely on developers recompiling -- which didn't happen for mobile and probably would never happen for this device either. Again, still waiting on proof it can do so.

I can provide you with a few links on Win32, but that's likely to lead to a debate as well, which for the topic at hand would be a distraction. We can return to that later.

Please. I think people would be interested. I've seen/heard nothing which even hints they can run Win32 apps. Given it isn't running Windows on ARM, and (as we agree) Win32 is virtually useless on anything under 10" with touch and often not even then if scaling is needed), seems like a lot of work for (presumably) only a continuum-mode.

 

[fatclue_98]

Casual IT users? Nope. In some corporate settings? Unfortunately... Yes... (most definitely if we take the 10 year mark).

My Cardiologist and Rheumatologist's offices both still use 2000 Professional so yes, there will be outliers. I was thinking more mainstream users. Heck, places like Home Depot were still using WinMo 6.5-based handheld inventory tracker/scanners as late as last year until Microsoft finally pulled the plug on WinCE support.