Yeah. My understanding and experience in the software industry typically leads me to far more accurate predictions than anything most bloggers might say. As we stated elsewhere, most of it is BS ;-) I think I've pretty much ignored everything on this issue except what I've seen from MS, but that has been very little.
Before I start, let me clarify my terminology: Windows Professional, Enterprise and Home are different SKUs. They are editions of the same OS. In contrast, I view W10M as a different "Windows flavored" OS. To call two operating systems the same, people with a CS background typically consider the ability to run the same software (without emulation) the main criteria. W10M doesn't run Win32 software while W10 can, which is why I consider them to be different OSes.
My background is in business computing, specifically networking, although I have done some basic coding in my time.
I don't think that is the standard technical view of OS, ie which APIs are included for running software. For example blackberry 10 features the android runtime, but I am not sure anyone would argue that its a flavour of android.
The entire software stack, right down to the hardware level, ie the OS as a whole, I think is how people technically consider an OS. In that respect, win10m and windows 10, are relative, they share one core, and will one day share cshell. They also share a lot of other code.
Technically the simple linguistic view of an operating system is "An operating system (OS) is system software that manages computer hardware and software resources and provides common services for computer programs.", which is perhaps something in between we see this, as software developers and network engineers respectively. However even on that level, win10m and windows 10 for the desktop share enough code that they provide similar hardware management and services, even if they handle software management differently.
Identity as a form of shared property possession, is probably a little philosophical however, and I'd betting some people would split it slightly different ways.
OK, but that doesn't necessarily make Andromeda a separate SKU or "relative" however. For example, right now we have at least two branches of W10M. One in the main Windows branch and the feature2 branch. Both are just plain old W10M though. They are merely separate development paths where different features are added or experimented with. In fact, this is standard MO for any project at MS, as each will typically use their own branch for development and eventually merge back into the main branch when things are stable enough for release.
My point here is only that an insider saying that Andromeda is developed in a "branch of W10M" doesn't really mean anything EXCEPT that Andromeda is derived from W10M (the Win32-less version of Windows).
If Andromeda ends up being the only branch of W10M that receives sustained developer attention, and it can run W10M apps, then that would make Andromeda the re-branded successor of W10M (like W10M was the successor of WP8).
True, but it doesn't mean win10m is being replaced either. The language branch only refers to being based on. The topic of whether win10m is getting further features as promised, or how long it will be supported, often ends up being speculative, subjective and pretty fruitless, and we could probably just dump that one to avoid that dumpster fire.
Judging OS "relatedness" by their UI is a highly inaccurate way of judging similarity. Take W10 for example. It can run on a 8" screen or on three 40" displays simultaneously. It can be used with a pen, a touch screen, a mouse and keyboard, or a VR headset. Those are also very different experiences. Going by your understanding those would be different OSes.
In this case, perhaps our different background, lead my language to be confusing. I was talking about the GUI. Not the user interface. Specifically the way this OS is working on the surface (running two related screens with separate apps, or one app with two separate pages of information, and the ability for those to interface, and dynamically shape themselves in response to magnetic connection orientation etc). The way you interact with the OS software on this device will be quite radically different from win10m.
Does that mean a separate SKU? Well, technically it might not, given some of that could all be packed into composable shell, and spread across every SKU identically, however the fact that Andromeda is listed in the composable shell code amongst all the other SKUs such as windows 10 home, and windows 10 mobile, suggests everything that I originally said - that it is a relative OS, and the win10 is going to continue as a separate entity (else why code cshell for it). That composable shell code is present in current APIs, in current windows code.
Perhaps they will change their mind, but this stuff isn't just being pulled out of a hat, I'm working on a wholistic deductive process here, matching data points together. It's not a lof of data, and I could be wrong, or they could change their mind, but this is how it looks to me.
Of course that doesn't specifically mean it can't run win32s either, although I don't see the point in that, personally. win32's spanned across a screen with a gap in the middle would be the user bugbear of a lifetime.
For various reasons, we just know that they are all powered by the same OS.
Well, IDK. I would call it a very closely related OS. Windows home for example doesn't have domain joining ability, so it differs on the network layer.
Finally, W10 also contained a lot of new code over W8. That didn't make it an entirely different/distinct OS either. Everybody would agree that W10 was the successor of W8 and the same OS.
I wouldn't call it the "same OS", any more than I'd consider 95 "the same OS" as windows 10. They are closely related OSes in my mind, which share a historical lineage. Sorry to split hairs!
Again, my point is that none of these issues are a good way of judging whether Andromeda is the successor of W10M or a separate OS. Judging by the UI is probably the worst possible way.
Well again, its none of these singular points that has lead me to this conclusion. Perhaps that's just hard to convey without a person seeing all these little things first hand. I don't think I have a complete picture, and I guess time will tell anyway.
I see no reason W10oA couldn't take advantage of GPS or always on LTE etc. Always-connected functionality has been a part of Windows ever since the Intel Atom. W10 already includes the same location services W10M does, so I see no reason why W10oA couldn't also support GPS. That claim makes little sense to me. Apart from that I agree with everything else you said about W10oA vs UWP.
I meant win32.
yes, win32 CAN take advantage of scaling, always on LTE, GPS and toast notifications. They generally don't but they could. This is true.
But at which point, running a native UWP on windows on arm would be even closer, in terms of the code, and getting that 30%+ extra performance would just seem more attractive. Running your software on an emulation layer for a significant portion of users, at reduced performance is never going to seem ideal to software companies, even if the software is very light. Whilst technically it may work okay, it will still work better as UWP.
Supporting or not supporting Win32 is pretty much the only thing that warrants maintaining a separate OS.
I am not privy to what services, GUI, networking qualities, or other features this new SKU possesses. MSFT is pretty experienced at programming, I am sure they know what they are doing though.
Andromeda being a separate Win32-less OS would run counter to that.
It seems like they are running against that in general these days. Windows S, windows on arm - they are increasing SKUs atm. I don't think that's a permenant thing, I think its a bridge/roadmap type thing, to encourage UWP, before later, everything win10m included is merged.
ATM the hardware isn't really up to everything from IoT core, being the exact same OS. With moore's law slowing, it could be awhile before we see the IoT getting 4gb of ram standard. To a certain degree, running shared code and services is as far as it can go right now (one core, cshell, networking later, Cortana etc).
Furthermore, MS gave up trying to popularize the app store directly. They decided they can't make it popular except through Win32 compatibility, which is why they put W10M on the back burner.
That's the whole point of W10oA. Creating another Win32-less version (not just an SKU) of the Windows OS would also run counter to that.
That's the exact opposite to how I see their current strategy and windows on arm. But I explained that in the last post, how windows on arm, is actually a way to reduce win32s,not encourage them. Much like a kind of backward compatibility, in order to promote the new thing.
Windows s encourages centennials.
windows arm encourages native UWP, and also apps that take advantage of built in as standard always connected LTE
Cshell encourages, along with WoA, app scaling.
It's a bit like sheep dogs, herding software developers
If MS intends to release another Win32-less version of Windows, then putting W10M on the back burner would have been the stupidest idea ever, as that has severely reduced interest from all parties.
Has interest in UWP dev actually decreased though? I see big software development houses doing UWPs, recently. Certainly developing specifically for mobile may have decreased. But there are advantages of doing full UWP, beyond mobile, even currently before cshell and windows on arm are released.
Not that MS doesn't ever do stupid things, but they aren't that stupid. This would basically represent another strategic shift before they even got around to implementing any part of the previous shift.
I don't see anything they are doing atm as stupid. The work on bugfixing win10m, and focusing of bugfixes makes a lot of sense in light of the notion of a branch thereof for a new product.
This is again just to explain why I remain skeptical of the idea that Andromeda will be a distinct OS alongside W10 (of which W10oA is just an SKU) and W10M. None of the technical reasons for it being distinct really hold any water, and strategically it just doesn't jive at all with MS' current strategy.
Everyone is entirely welcome to believe what they like. Interesting how differently we view MSFTs current strategy actually though. I see everything right now, as herding people, piecemeal into UWP, you see them as moving back to win32, and sort of giving up on their universal, multi-platform approach? Is that right?
Despite knowing nothing about Andromeda, MS' strategy and the technical framework they operate from suggests it must be based either on W10oA or it must be the successor to W10M (both supporting novel hardware).
I don't really see that.
I can't in fact see any real benefit to even running win32 on a mobile sized screen. It's horrible enough an experience on an eight inch tablet. Bringing that experience to an even worse UX level, would just seem to be terrible strategy. Continuum might be "a benefit", but continuum isn't even particularly popular as a feature. Its more for enthusiasts. Even with refinement I think it'll be a long time before it catches on.
WoA is different because at least its a bigger screen.
I also don't really see the benefit in wildly changing win10m. It could use some minor graphic tweaks, some bug fixes. And of course, ongoing feature support. But there's nothing actually wrong with it, conceptually.
Perhaps Silverlight support should be dropped, to further encourage UWP, when the timing is right. And cshell makes sense, but that's a pretty minor GUI change for the most part, and its supposed to be coming to every single SKU - Would that then make windows home "rebooted", and IoT enterprise? I don't think so.
I can't think of anything substantial that either would be changed or needs to be changed right now. To add something like arm win32 emulation, you'd actually need a common use case scenario - a graphene based screen (years off), a table projector (maybe, but its a battery drainer, and is still somewhat niche). Plugging a phone into a dock, just isn't something I see a lot of people doing no matter how good continuum is, unless they lack money for multiple devices.
So yeah, personally I can't really see the point of win32 on a phone, nor can I see what would be, or needs to be 'rebooted' in windows mobile. Minor graphical tweaks and bugfixes aside (I don't think those count as reboots".
If you really want to interpret "enterprise focus" as meaning "features for enterprise users" (which I don't think matches MS' interpretation), then you probably shouldn't take that to mean multitasking and task switching. Enterprise features are things like share-point integration, manageability, azure authentication, etc.
IMHO the ability to work with multiple windows or run background processes is a productivity feature, which anyone who has ever used a desktop computer simply expects from a device that facilitates desktop like usage scenarios. Selling that as an enterprise feature would probably get MS laughed out of most CTO's offices.
Okay, your probably technically correct. Although clearly productivity features will sell better to enterprises, professionals and so on than every day consumers.
Let's just say "multi-tasking is not a feature that is going to drive the average joe consumer to buy this new device". It's a niche product.