Linked by Thom Holwerda on Thu 18th Oct 2012 11:58 UTC, submitted by lucas_maximus
Windows "This may be a good sign for Microsoft: a little over a day after putting its new Surface RT tablet up for pre-order, the entry-level $499 version of the tablet has sold out. Its estimated shipping time has slipped from October 26, Windows 8's release date, to a more nebulous 'within three weeks'." We'll see. Wouldn't be the first time a company artificially keeps supply short to generate 'sold-out' hype.
Thread beginning with comment 538969
To read all comments associated with this story, please click here.
Maybe not this time.
by przemo_li on Thu 18th Oct 2012 12:48 UTC
Member since:

MS limited "numbers" for sale to 3 mln this year.

And lets face it, its THE WinRT tablet. Anyone who want to have Win on ARM (why is beyond me ;) ) will go to pre order.

I think that numbers will be below 100k. Like 10% of 1 mln. Like 3 months left of this year with 3mln year sales as MS target.

Reply Score: 0

RE: Maybe not this time.
by some1 on Thu 18th Oct 2012 15:20 in reply to "Maybe not this time."
some1 Member since:

No, this is actually THE Windows RT tablet. WinRT stands for Windows Runtime:
Welcome to the new Microsoft naming strategy.

Reply Parent Score: 3

RE: Maybe not this time.
by DOSguy on Thu 18th Oct 2012 16:36 in reply to "Maybe not this time."
DOSguy Member since:

why is beyond me

the break with win32 on RT might prove to be a very good move in the long run. Devices where battery preservation is critical, are not served with lazily ported Windows applications.
Yes, this means there will not be a large number of apps in the beginning, but if these tables sell reasonably well this won't be a problem for very long.

Reply Parent Score: 2

RE[2]: Maybe not this time.
by Alfman on Thu 18th Oct 2012 17:05 in reply to "RE: Maybe not this time."
Alfman Member since:


"the break with win32 on RT might prove to be a very good move in the long run. Devices where battery preservation is critical, are not served with lazily ported Windows applications."

Do you have evidence that battery preservation is a problem with win32s in the first place? It's a serious question since I've never seen anyone back this assertion with real data. At their core, win32 apps are event oriented, meaning that they generally only consume CPU when they're interacted with. Unless a developer has used a bad practice of polling, it's not at all clear to me how switching the API will help save the battery.

I might believe the possibility that bad practices are rampant on win32s, however even there I'd have to question this argument's validity since the win32 software I use rarely consumes CPUs in the background unless it's a daemon designed to do so.

If you're talking about foreground apps, there's nothing really preventing a metro app from wasting CPU.

If you're talking about background apps, there's no reason the OS could not have a policy to completely suspend or significantly reduce CPU to win32 apps in the background. Well behaved win32 GUI apps would not be affected during background suspension. I'd even expect the majority of "poorly" behaving win32 apps (like a ray tracer) to wake up and resume in the foreground without a hiccup.

My point is, I don't think there's any overwhelmingly strong technical reason a power policy for metro apps could not be applied towards win32 ones. And I've never seen any evidence to show that the new APIs are somehow more efficient. I don't want to assert that they are not, but I definitely need to see solid data before accepting that they are.

Reply Parent Score: 2

RE: Maybe not this time.
by Nelson on Thu 18th Oct 2012 16:57 in reply to "Maybe not this time."
Nelson Member since:

I agree about Windows on ARM being useless. I just don't see the long term advantages.

Look at Intels roadmap over the next few years and you'll see them rapidly scaling down power consumption and scaling up GPU performance to erase ARM's advantage.

Intel's SoC designs also incorporate connected stand-by which means that Win8 tabs with those chips will have decent battery life. If they can get even eight hours out of a charge then its already mostly close enough. (Note: This applies to Atom SoC designs only, not high end Win8 tabs with Core i7's)

I don't think Win32 will go away fast enough for Windows RT to outsell Windows 8. I think that in the long run, people will have their tablets be the center of their digital life and just dock them to output to a monitor, and use Bluetooth keyboards and mice to get productive stuff done.

Intel will win the architecture war. I have full faith in their expertise and their leads in fab technology.

Reply Parent Score: 2

RE[2]: Maybe not this time.
by EvilMonkeySlayer on Fri 19th Oct 2012 15:33 in reply to "RE: Maybe not this time."
EvilMonkeySlayer Member since:

Intel should be able to get some market share with their mobile processors, but they've got a snowballs chance in hell of becoming the dominant processor architecture for mobiles.

In order for them to win they'd have to convince the manufacturers of phones/tablets to limit themselves to a single manufacturer (since Intel own the x86 arch, to my knowledge the only others allowed to make x86 arch processors are AMD and VIA). So, they'd limit themselves to giving a monopoly to Intel where they can price gouge them versus the very competitive ARM market where you've got various fabs and manufacturers competing against each other making ARM processors.

Reply Parent Score: 3

RE[3]: Maybe not this time.
by zima on Thu 25th Oct 2012 23:58 in reply to "RE: Maybe not this time."
zima Member since:

Intel roadmap included, not a long time ago, also 10 GHz Netburst CPUs. And in ~embedded, Intel was supposed to "erase ARM's advantage" one or two times already...

Thing is, ARM doesn't stand still, and has many players heavily invested in it, and will offer more integrated, tailored to every imaginable usage scenario (it's about vast number of manufacturers) & less expensive solutions.
Intel can't and won't provide this, process lead is not so significant when you want to have low cost (where older processes are fine) with simply adequate performance, limited mostly by screen and radio module vs battery tech.

And there's more than ARM, also MIPS for example...

BTW, in early Intel presentations of Atom SoC there was a block cryptically named ~"32bit RISC system controller" in the ~southbridge ...I wonder what that was. Also, Infineon (acquired by Intel) radio modules certainly still use ARM cores.
Which means that in an "Intel inside" mobile phone there very well might be more ARM cores than x86 ones.

Reply Parent Score: 2