What's new

DisARMed: Microsoft kills Windows RT - Surface 2 - What went wrong

GreyFox7

Super Moderator
Staff member
Really it was dead before it started, if given a chance it could have been so much more, but here's what one Analyst says:

http://www.extremetech.com/computin...-kills-windows-rt-future-arm-support-doubtful

One of the most striking reversals over the past three years has been Microsoft’s embrace and subsequent backpedal from the ARM ecosystem. In 2012, when Microsoft first demonstrated Windows on ARM (later named Windows RT), the technology world was in broad agreement that this bold move to support a non-x86 architecture would be a vital component of future Windows devices. Today, it looks like that support is dead or at least on life support. At its Windows 10 event earlier this week, Microsoft confirmed that there will be no Windows 10 for current users of Windows RT. Windows RT users will instead receive some unspecified “feature” updates.​

In the time since Intel has dumped Billions into subsidies for its components in small devices (phones and tablets) without which it's unlikely there would be a single small device using their platform. Despite this ARM still rules on Phones and arguably Tablets as well. At least the ones that users don't complain about heat problems. If there was a knock on ARM it's that they were underpowered CPU wise. The knock on Intel's entries was they were underpowered GPU wise and drew too much Power in watts. Intel has gotten close but still lags in these critical areas but also in total SoC features requiring more parts per system. ARM on the other hand has pushed the envelope on GPU designs and done well on the CPU front also. The current iterations in early 2015 show promise of being very competitive in the CPU performance compared to the best low powered (watts) Intel SoCs.
From this we can not conclude that the ARMs race is over but without Intel subsidies it likely would be.

Windows RT: Maligned and mismanaged
The non-adoption of Windows RT was a perfect storm of poor decisions at multiple levels of Microsoft. The Tegra 3 SoC that powered the first generation Surface RT didn’t pack enough horsepower to really drive a top-tier tablet experience. Microsoft did a miserable job communicating the differences between Windows RT and traditional x86 Windows, tales spun of consumers taking the hardware home, firing it up, and then returning it when they realized they couldn’t run desktop applications.
Yep, yep, yep, yep.

In retrospect, Microsoft’s decision to retain the desktop for Windows RT was a huge mistake. While it’s true that the desktop offered certain functions that the nascent Metro design wasn’t ready to replace on launch day, the sight of familiar icons and desktop interfaces primed consumers to expect the same degree of software compatibility and flexibility. Stripped of those features, Windows RT lost much of its appeal.​

Microsoft's implementation of the RT Desktop was necessary. Preventing developers from using it for their apps and trying to force them into Metro was the mistake. They would have figured out how to use the Desktop UI and realized the benefits of Metro but forcing their hand was totally off putting and it still is. Lord knows you cannot force a developer to do anything they only work if they think they have total control and freedom to do as they please. Clearly Excel, Word, PowerPoint, OneNote, and all the standard Windows tools demonstrate the Desktop is viable for an ARM compiled app so imagine the devs dismay when they find you cannot go there with your own ARM app. We will get to how x86 Apps could have been handled in an ARM environment later.

More to follow...
 
Last edited:
OP
G

GreyFox7

Super Moderator
Staff member
Continued...
The entire assumption of Windows RT, the reason it existed is because Microsoft bet that Intel would be unable to deliver a CPU that could compete with ARM devices in the lower end of the tablet market.​

Arguably this is still true, yes there are plenty of low end tablets with Intel SoCs but would there be without billions in subsidies? Conversely ARM is approaching i5 performance levels while maintaining the lead in GPU performance and completeness of feature set. In 2015 the ARM family is more potent than ever.

The initial launch price on x86 tablets reflected this assumption — Microsoft plus Intel would rule the high end of the tablet market, while ARM-based Windows 8 products would take over the lower end. Unfortunately, it quickly became clear that the ARM products in question were going to completely fail to uphold their end of the equation thanks to a crippling lack of software and a bad overall experience.
ARM or ARM products did not fail here. What failed was the strategy to get APPs onto the platform handcuffed by Microsoft's arbitrary decisions to force Metro only and prohibit developers from building Desktop apps which undermined the developer community and sabotaged the platform from the start.

Meanwhile, Intel kept iterating, eventually delivering tablet and smartphone processors that could match their ARM counterparts in ways that previous generations of hardware couldn’t accomplish.
Meh, still lagging although getting better. Full Windows and regular Windows Apps poor quality creates stability issues which frankly is unsuitable for the optimal Tablet User Experience. There are ways to address this but its not being done.

It’s been clear for several years that Microsoft had no idea how to fix RT. Microsoft updated the hardware once, in the Surface 2, but it never really revisited the bad set of assumptions that handicapped the device.

It would’ve made little sense to ship an x86 emulator on the original Surface RT, but a hypothetical third-generation RT device (if one existed) would likely be based on Nvidia’s Tegra K1 or an equivalent 64-bit chip from Samsung or Qualcomm. Devices like these would’ve been powerful enough to run x86 software, even if they didn’t run it particularly quickly, and could’ve eventually bridged the software gap.
Yes, absolutely an x86 emulator on current high end ARM SoCs is viable. In addition has anyone heard of the Cloud? x86 Apps as a service is totally doable and has been since DAY ONE. Upload your app to the cloud and run it remotely would work now and it would have worked three years ago.

Instead, Microsoft is going to focus on its Windows Phone devices and grow its ARM support in that product line.
Pfft.

There's an Elephant lurking in the room which I wont bring up now... maybe later.
 
OP
G

GreyFox7

Super Moderator
Staff member
Apple is rumored likely to be going to the ARM platform for lower end MACs while MS is falling in line with Intel. Will all MS Phones soon be Intel also???

Abandoning or even contracting ARM presence is a big mistake IMO. Things are not the same as in the 90s when Intel crushed Alpha, MIPS, and PowerPC. It's too early in this contest to assume Intel has it won.

There's a mountain of software ready to step up to the next gen ARMs and its just an incremental evolution. Look back to the early Windows era when they were the challenger to the dysfunctional and incompatible world of Unix and Mainframes. It was the army of apps and developers that made it all possible and the uniformity of Windows PCs that enabled the apps and developers to out hustle the incumbents and heirs apparent with Windows and Intel PCs. The roles are reversed now but the battles are not decided. Id submit that Windows Desktop software ecosystem has a tougher road ahead but platform flexibility would help it break free from the past and move forward.
 

jnjroach

Administrator
Staff member
I think that the Windows Market for large ARM based tablets wasn't large enough to keep the line, which kills me as I loved my Surface 2. Also Windows users are very reluctant to move away from Win32 Applications.

I do believe we will see a 7 or 8" ARM Tablet from Microsoft under the Lumia Branding at the Windows 10 launch, and I also think we will see N-trig on it and maybe even on the Phablet (which has a bigger market).
 
OP
G

GreyFox7

Super Moderator
Staff member
For me that 7-8"er wont matter as my preference is not to use anything smaller than 10.1 unless its to make a phone call, send a text, or short email.

If its a choice between my phone and hauling out my laptop I'll do more on the phone but as soon as I have the option to work on a decent sized tablet I'm more inclined to use that for anything more than the basics. Although a Pen on a phone/phablet will get it used more often for more things but size does matter.

The 6 to 9.9 inch range is "no man's land" aka The no GreyFox Zone.
 

jnjroach

Administrator
Staff member
Phablets are still a growth market (only a few players)....large tablets (ARM) have plateaued and small tablets are decline.
 
OP
G

GreyFox7

Super Moderator
Staff member
Yeah, I suppose people are not upgrading like they used to with some devices and maybe we haven't determined what level of upgrade would entice a tablet user to upgrade although from Tegra 4 to Tegra X1 or Snapdragon 810 is a significant performance increase. However if all you can do is browse the web and run the same Apps it may not matter even then. :)

I think more built-in features will be needed not just a new SoC.
See Surface Pro 4 - wishlists!
 

oion

Well-Known Member
For me that 7-8"er wont matter as my preference is not to use anything smaller than 10.1 unless its to make a phone call, send a text, or short email.

If its a choice between my phone and hauling out my laptop I'll do more on the phone but as soon as I have the option to work on a decent sized tablet I'm more inclined to use that for anything more than the basics. Although a Pen on a phone/phablet will get it used more often for more things but size does matter.

The 6 to 9.9 inch range is "no man's land" aka The no GreyFox Zone.

Yeah, I'm not going under 10 inches for a usable tablet.

I had written my own reasoning for the Surface RT market problems, but don't intend on updating until we get a clearer picture of what we'll get from Win10 and maybe universal apps. The whole line had too many meta problems, though the product itself isn't bad. Universal apps would be great, but like I said in another thread, a consumer doesn't buy a device thinking "if this doesn't support the next generation OS, it's crap."

Meanwhile, there are some crazy deals going on for old Surface gens. Hmm.
 
OP
G

GreyFox7

Super Moderator
Staff member
There was a deal on the Surface 2 but it seems to have ended perhaps because there aren't any :)
 
OP
G

GreyFox7

Super Moderator
Staff member
ARM twisting is one way of describing it. :)
Subsidies would be another..
"Over the last two years, Intel's mobile chip division has lost $7 billion while heavily subsidizing the manufacturing costs of Android Atom tablet makers. It now plans to phase out those generous incentives"
http://appleinsider.com/articles/14...ly-subsidizing-cheap-x86-atom-android-tablets

http://www.zdnet.com/article/intel-to-hit-40-million-mobile-chip-units-with-aid-of-subsidies/
"What's going on? Intel is subsidizing its chips — practically giving them away to partners — and recording the transactions as "contra revenue." Intel is shipping its mobile processors and offering refunds or rebates to the vendor for using them. "

"Going into 2015, Intel's latest mobile chips won't carry the subsidy"
 
Top