Windows CE Update Set for Mid-2003 Release

[Jason Dunn]

<div class='os_post_top_link'><a href='http://www.infoworld.com/articles/hn/xml/03/01/13/030113hnwinceupdate.xml?s=IDGNS' target='_blank'>http://www.infoworld.com/articles/h...ate.xml?s=IDGNS</a><br /><br /></div>"The next update to Microsoft's Windows CE .Net OS, which has been given the code name McKendric, will ship before the middle of 2003 and include several new features, a company executive said at the Microsoft Embedded DevCon 2003 conference being held here. Expected to be called Windows CE .Net Version 4.2 when it hits the market later this year, McKendric offers several improvements over the current version of the OS, Windows CE .Net 4.1. It will incorporate new features, such as support for VoIP (Voice over Internet Protocol), a firewall, and wireless networking support for residential gateways, said Chip Schnarel, group program manager for Windows CE OS development at Microsoft.<br /><br />Among the changes being made to the OS is a feature designed to improve the compatibility of APIs for software developed for PDAs (personal digital assistants) running Microsoft's Pocket PC software and other devices running software based on Windows CE .Net, Schnarel said. To make this possible, McKendric will allow software developers to change an application's user interface so that it can be displayed on differently sized screens while maintaining compatibility at the API level, Schnarel said. This would allow developers to more easily port their applications to different Windows CE-based devices, he said."<br /><br />Due to a Microsoft NDA I'm under, I can't say much about this news, but if you read between the lines you can see this is going to be a very exciting year for the Pocket PC!

[Jonathon Watkins]

Due to a Microsoft NDA I'm under, I can't say much about this news, but if you read between the lines you can see this is going to be a very exciting year for the Pocket PC!

Good man. Just let us know when you can. :D

That screen re-sizing sounds good - roll on 640x480 transflexive screens. :P

[mclaughlinc]

I just hope that my new iPAQ 5455's 48 mb of ROM is big enough to hold the new version when it comes out.

[Seraph1024]

I am waiting.

[JMountford]

Ya gotta love NDAs....

Well anything you can say or at least point us in the right direction is appreciated.

[Jimmy Dodd]

Has MS actually mentioned whether or not we will see a new Pocket PC version at that time?

[FredMurphy]

We appreciate that you can't spill the beans but surely it's gotta be time that Microsoft want to disclose "Pocket PC 2003"..... :?

[WillyG]

Im trying to read between the lines here.

Open question 1: Will the new Pocket Pc version be based on (the newest) Windows CE .Net (Version 4.2), or an older version?

Open question 2: If its based on 4.2, when could we expect it out? Let me try an optimistic guess....Cristmas 2003(?), another moth or two for hardware vendors, do do their work and make it availiable for download. Then we have it flashed (if you succed ;)), somwhere in January/February 2004..........(!)
But if its based on an elder version, we could have it by summer i guess.

Open question 3:
What would the new ship be named? Pocket PC 2003? 2004? or somthing else? *remembers the earlier post on .NET name use*

[sponge]

I'd be very surprised if 48MB couldn't do it. I do think we'll be seeing what we saw with the 3630 to PPC2K2, as much as I don't want to see it happen :\

[FredMurphy]

All the more confusing with devices coming out running CE.NET without a Pocket PC OS.

Argghh..... dammit! Who knows and can be bribed? :o Someone must have malleable priciples. We want to know!! Name your price. :lol:

Fred

[heov]

I hate the .NET name, as I do w/ the year names...

Why can't they just stick to version numbers... Like PPC 2.3 for PPC 2k2 SP3?

You might say because it's better for consumers, but they're really not that dumb to differentiate a product w/ a year and version number...

Let's hope they don't start calling it "Pocket PC .NET 2003"

[Rob Borek]

Im trying to read between the lines here.

Open question 1: Will the new Pocket Pc version be based on (the newest) Windows CE .Net (Version 4.2), or an older version?

Open question 2: If its based on 4.2, when could we expect it out? Let me try an optimistic guess....Cristmas 2003(?), another moth or two for hardware vendors, do do their work and make it availiable for download. Then we have it flashed (if you succed ;)), somwhere in January/February 2004..........(!)
But if its based on an elder version, we could have it by summer i guess.

Open question 3:
What would the new ship be named? Pocket PC 2003? 2004? or somthing else? *remembers the earlier post on .NET name use*

All 3 of those open questions will be answered when (and if) Microsoft issues a press release announcing the next version of the Pocket PC OS...

[Jonathon Watkins]

All 3 of those open questions will be answered when (and if) Microsoft issues a press release announcing the next version of the Pocket PC OS...

Thank. That's so helpfull, yet at the same time..... it's NOT. :wink:

[jdhill]

I have no information on this, however, I do remember seeing a reference to Pocket PC 2003 in a Microsoft document recently. Therefore, I think that we can safely assume that it's name will indeed be "Pocket PC 2003" and that it will be released this year.

As to what version of Windows CE it will be based on, I'd hazard a guess that it will be Windows CE .NET 4.1 rather than the just-announced 4.2.

[Foo Fighter]

Due to a Microsoft NDA I'm under, I can't say much about this news

Oh come on! NDA violations are no big deal. So what if they sue you for 22 million...or send you to prison for a couple years. That's nothing. :wink:

[pgh1969pa]

Come one guys, reading between the lines is not difficult....my read is we will be recieving higher resolution screens this year with the new OS. Also, I would guess that WinCE 4.2 will be at the core of PPC 2003. There's been a couple of folks on the web that imply that the new PPC OS is due mid year.

[Jason Dunn]

We appreciate that you can't spill the beans but surely it's gotta be time that Microsoft want to disclose "Pocket PC 2003"..... :?

Nope. Microsoft doesn't talk about their Pocket PC OS releases until very shortly before the hardware starts shipping. They have no desire to kill OEM sales - look how long people were waiting for the fabled iPAQ 54xx devices....what, eight months of rumours and speculation? It might be fun for us, but it hurts the industry.

[Sslixtis]

I hate the .NET name, as I do w/ the year names...

Why can't they just stick to version numbers... Like PPC 2.3 for PPC 2k2 SP3?

You might say because it's better for consumers, but they're really not that dumb to differentiate a product w/ a year and version number...

Let's hope they don't start calling it "Pocket PC .NET 2003"

I'd bet there are lots of consumers that dumb! There are still consumers out there that think all Handhelds are Palm Pilots. :shocked!:

As for naming conventions, I could live with Pocket PC 2004, but agree anything with .Net in it would be "bad".

We have got to be coming up on the 18 month life cycle of the Handheld OS soon though. Wasn't Pocket PC 2002 released in September/October 2001? :eek:

[donkthemagicllama]

Nothing mentioning it being optimized for XScale?

I'm not buying a new device until the performance issues get straightened out, whether it's a faster data bus or an optimized operating system.

Although the fact that it's hinting at resolutions greater than 320x200 is good. That's where PPC is starting to lag behind PalmOS (at least Sony's devices).

[Foo Fighter]

Although the fact that it's hinting at resolutions greater than 320x200 is good. That's where PPC is starting to lag behind PalmOS (at least Sony's devices).

This may also explain why PPCs based on current technology are getting cheaper. 320x480 may be the future of high-end Pocket PC hardware, with 320x240 dedicated to low-end hardware.

[van_mierlo]

Although the fact that it's hinting at resolutions greater than 320x200 is good. That's where PPC is starting to lag behind PalmOS (at least Sony's devices).

This may also explain why PPCs based on current technology are getting cheaper. 320x480 may be the future of high-end Pocket PC hardware, with 320x240 dedicated to low-end hardware.

Don't forget that there are already high resolution screens available on the Pocket PC 2002 operating system......Some Tablet pc's

[brntcrsp]

Although the fact that it's hinting at resolutions greater than 320x200 is good. That's where PPC is starting to lag behind PalmOS (at least Sony's devices).

This may also explain why PPCs based on current technology are getting cheaper. 320x480 may be the future of high-end Pocket PC hardware, with 320x240 dedicated to low-end hardware.

Tell that to HP.

[Marcel_Proust]

Although the fact that it's hinting at resolutions greater than 320x200 is good. That's where PPC is starting to lag behind PalmOS (at least Sony's devices).

This may also explain why PPCs based on current technology are getting cheaper. 320x480 may be the future of high-end Pocket PC hardware, with 320x240 dedicated to low-end hardware.

Don't forget that there are already high resolution screens available on the Pocket PC 2002 operating system......Some Tablet pc's

Well really the HPC 2000 series is what you mean. The Jornada 728 has recently been discontinued - I wonder if the replacement for it is one of these? But that would be too soon.

[Daniel]

Hmmm VoIP + Windows CE .Net 4.2 = Neonode with extas. :)

There are some very intersting ideas to come out of this. I guess I'll just have to wait for this before I buy anything (hmm, could be saying that for a while though :().

I'd love to see PPC2k3 based on WCE.NET 4.2 (please Microsoft change that to Windows CE 2003!!) but it wouldn't seem likely unless the two teams have decided to release version at or around the same time. Normally there is a fair amount of time between WinCE and PPC releases. Perhaps the whole "read between the lines" thing was meant to say that because the APIs were going to be changed such that they work on all PPCs & CE based devices then Windows CE.Net 4.2 would be our next PPC OS? In which case PLEASE PLEASE PLEASE rename it!!! ;)

Daniel

[Ekkie Tepsupornchai]

This may also explain why PPCs based on current technology are getting cheaper.
May have had something to do with it, but consider that of the cheaper PocketPCs, the HP1910 and V35 (while still relatively great values compared to previous devices) definitely cut some corners to cut down on the cost. I'm pretty convinced that Dell would be attacking the typical PPC price ranges regardless of whether a new PPC was on the horizon or not.

320x480 may be the future of high-end Pocket PC hardware, with 320x240 dedicated to low-end hardware.
I get the feeling that this new OS is going to allow vendors the ability to create a wider range of devices supporting all sorts of resolutions.

I can see low-end "PPC2003" devices at the end of the year sticking to 320x240 while high-end devices go up to 480x320 and the OS is probably flexible enough that when the vendors are ready, they can go up to 640x480 and still have full OS operability.

Also, I wouldn't be surprised if the OS can now handle landscape modes with this new resolution flexibility.

BTW Foo, wasn't it you that just recently predicted that the PPC OS will not support more than 320x240 anytime soon? :wink:

[paris]

Well as i understand it one thing is sure for the furure of pocket PCs

OS SUPPORT FOR ANY ORIENTATION
OS SUPPORT FOR ANY SCREEN SIZE
=> APPLICATIONS TO SUPPORT ANY ORIENTATION + ANY SCREEN SIZE

;)

[R K]

Uh oh... I read between the lines and I just saw an iPAQ with an integrated foldable keyboard coming.
This leads me to guess that the next PPC OS will use Win CE 4.2, plus since Jason is running a PPC site and tells us to read between the lines on this particular article, I think my guess has some backing.

Let me say that again. I smell an iPAQ with integrated foldable keyboard running WinCE 4.2 with the next Pocket PC OS coming down the road.

Now, I'll run with the 18 month OS life cycle thing and guess a release of May - June, 2003.
Only time will tell.

[Ekkie Tepsupornchai]

This may also explain why PPCs based on current technology are getting cheaper. 320x480 may be the future of high-end Pocket PC hardware, with 320x240 dedicated to low-end hardware.

Tell that to HP.

LOL... HP certainly hasn't flinched yet at Dell's presence... even the 39xx hasn't dropped much in price... sheesh, even the much inferior 38xx that are still in retail are selling for higher than the Dell devices.

[Ekkie Tepsupornchai]

Let me say that again. I smell an iPAQ with integrated foldable keyboard running WinCE 4.2 with the next Pocket PC OS coming down the road.

Now, I'll run with the 18 month OS life cycle thing and guess a release of May - June, 2003.
Only time will tell.

Introduction of better new devices might make me real glad that I held tight to my 38xx during this recent run of new PPCs... of course I wonder at the prices of these new gadgets.

I'm not sure I like the idea of an integrated foldable keyboard... unless it's detachable, it would hurt the overall portability of the device itself.

[Will T Smith]

Although the fact that it's hinting at resolutions greater than 320x200 is good. That's where PPC is starting to lag behind PalmOS (at least Sony's devices).

This may also explain why PPCs based on current technology are getting cheaper. 320x480 may be the future of high-end Pocket PC hardware, with 320x240 dedicated to low-end hardware.

Don't forget that there are already high resolution screens available on the Pocket PC 2002 operating system......Some Tablet pc's

Nope. These are based on last years CE.Net OS. PocketPC2002 is a branch from an earlier code base.


Needed features:

1) X-Scale optimization. Absolute must. Hardware vendors won't settle for less when their pimping 400Mhz devices that run at the same speed as the 200Mhz devices.

2) Support for up to 1Gig of native RAM. PocketPC needs some headroom for "Power Devices".

3) Support for up to XGA resolution devices. This is entails more than just hardware. The infrastructure MUST support a dynamically sized interface window. That is visual components must be specified with relation to screen size, not absolute pixels. That way apps will work on high end and low end devices.

4) OS driven support for redefining the screen orientation.

5) OS support for Firewire and USB On-the-Go

6) Native Bluetooth and Wi-Fi support.

7) Native DirectX for PocketPC. A Comprehensive Graphics and multi-media libarary so that vendors will include 3D hardware.

8) USB hosting for connectivity with mass storage devices.

[guinness]

Although the fact that it's hinting at resolutions greater than 320x200 is good. That's where PPC is starting to lag behind PalmOS (at least Sony's devices).

This may also explain why PPCs based on current technology are getting cheaper. 320x480 may be the future of high-end Pocket PC hardware, with 320x240 dedicated to low-end hardware.

Don't forget that there are already high resolution screens available on the Pocket PC 2002 operating system......Some Tablet pc's

Nope. These are based on last years CE.Net OS. PocketPC2002 is a branch from an earlier code base.


Needed features:

1) X-Scale optimization. Absolute must. Hardware vendors won't settle for less when their pimping 400Mhz devices that run at the same speed as the 200Mhz devices.

2) Support for up to 1Gig of native RAM. PocketPC needs some headroom for "Power Devices".

3) Support for up to XGA resolution devices. This is entails more than just hardware. The infrastructure MUST support a dynamically sized interface window. That is visual components must be specified with relation to screen size, not absolute pixels. That way apps will work on high end and low end devices.

4) OS driven support for redefining the screen orientation.

5) OS support for Firewire and USB On-the-Go

6) Native Bluetooth and Wi-Fi support.

7) Native DirectX for PocketPC. A Comprehensive Graphics and multi-media libarary so that vendors will include 3D hardware.

8) USB hosting for connectivity with mass storage devices.

Why not just use a laptop? The line between the 2 is beginning to blur, IMO. The prices for laptops are begining to fall, they have keyboards, built-in USB, many with large, hi-res screens and some come with WiFi. I know that portability is an issue, but if people would want all that wouldn't a laptop be more flexible? I would much rather have a laptop for $800 than an iPaq for the same price, I could type Office docs, burn cd's, watch movies and not have to suffer with a smaller screen.

[Timothy Rapson]

Needed features:

1) X-Scale optimization
2) Support for up to 1Gig of native RAM. PocketPC
3) Support for up to XGA resolution devices.
4) OS driven support for redefining the screen orientation.

5) OS support for Firewire and USB On-the-Go

6) Native Bluetooth and Wi-Fi support.

7) Native DirectX for PocketPC. A Comprehensive Graphics and multi-media libarary so that vendors will include 3D hardware.

8) USB hosting for connectivity with mass storage devices.



I nominate Will T. Smith for president....no ruler for LIFE, of the PDA universe.

[Foo Fighter]

BTW Foo, wasn't it you that just recently predicted that the PPC OS will not support more than 320x240 anytime soon? :wink:

Yeah, I was afraid that prediction might come back to haunt me. :wink:

[sponge]

I hate the .NET name, as I do w/ the year names...

Why can't they just stick to version numbers... Like PPC 2.3 for PPC 2k2 SP3?

You might say because it's better for consumers, but they're really not that dumb to differentiate a product w/ a year and version number...

Let's hope they don't start calling it "Pocket PC .NET 2003"

What, you mean Microsoft Direct XP CE.net 2003 Service Pack 3 is an excessive name?

[johncj]

OK, folks, let's clear this up:

Tablet PC's run Windows XP Pro (not embedded XP, the whole desktop OS + the extra tablet stuff) [1]

Smart Displays run Windows CE .NET 4.1 [2]

Pocket PC's run Windows CE 3.0

[1] http://www.microsoft.com/WindowsXP/tabletpc/evaluation/features.asp

[2] http://www.microsoft.com/windows/embedded/news/smartdisplays.asp

[Jason Dunn]

Just to be totally clear, I wasn't saying the next Pocket PC OS will be based on 4.2. It's reasonable to guess it will be based on 4.1 or 4.2, but I can neither confirm nor deny which one it will be. I feel like such a politician. :lol:

[Foo Fighter]

Sort of off-topic, but I really hope Microsoft drops the DOT NET naming scheme for CE as well. This whole Windows.net thing was rather dumb. :roll:

[ppcsurfr]

Wow... All the excitement... and then allthey do is release the compact framework for PPC and then call it PPC2003.

Hmmm. Keep it exciting guys... I just love reading all these speculations/predictions... care to trn this into another contest Jason?

ppcsurfr

[ExtremeSIMS]

&lt;snip>
Needed features:

1) X-Scale optimization. Absolute must. Hardware vendors won't settle for less when their pimping 400Mhz devices that run at the same speed as the 200Mhz devices.

Yes, still bus speed issues, but heck, I'll take this.

4) OS driven support for redefining the screen orientation.

Yup.

5) OS support for Firewire and USB On-the-Go

You notice the new ATI Imageon 3200 chipset supports USB OTG? Hmmm.... Reading between the lines and stacking it with ATI info:


12/15/16/18 bit TFT
Maximum resolution 320x480 at a color depth of 16-bpp
Partial display refresh
2 HW cursors, 1 HW icon/overlay
Frame modulation
Panel rotation (90°, 180°, 270°)
IMAGEON? 3200 is the first integrated co-processor to offer USB OTG
Connect video peripherals (such as a camera) to the handheld device
Supports CCiR 656, YCrCb 4:2:2 with 8-bit data bus
Capture resolution up to 640x480 pixels
SD Memory and I/O Controller


Rotation within the OS, USB OTG, 320x480 resolution, and possible video connectivity native?
Darn, and I JUST bought my Pocket PC. I guess I am getting wicked good use of it, though, so maybe I'll need to upgrade and give my wife the Toshiba e740 when this all happens. ;)

[seanturner]

I could be completely wrong, stoned, and wasted but, didn't MS say they weren't going to optimize Win CE for any processors? Although why they would do that makes no sense at all to me...

[seanturner]

Doesn't MS do all wince testing in house?

Does anyone else think there is a link between this and their PPC operating systems not being as stable and bug free as their desktop OS's. (By desktop OSes I don't mean me/98, those don't count, they're too much self riteous parasites)

[Sslixtis]

I could be completely wrong, stoned, and wasted but, didn't MS say they weren't going to optimize Win CE for any processors? Although why they would do that makes no sense at all to me...

They already customize for specific processors...ARM.

The problem right now is that it is ARMv4 and Xscale is ARMv5. And before Jason jumps all over me, let me add that some people claim the OS would be fine with the new Xscale processors if Intel didn't do such a lousey job with ARMv5 (couldn't be MSs fault, heaven forbid). Also, other semi-conductor companies do make ARM chips as does Intel, so we would be more likely looking at ARMv5 optimization and not Xscale optimization per se. However, to date I don't exactly see anyone really using Texas Instruments or Samsungs ARM chips but some out there are hopeful that the claims that the aforementioned are making about future chips will eventually come to fruition. :lol:

I don't care who makes it as long as it is FAST. The faster the better...W?BIC! :wink:

[Ekkie Tepsupornchai]

I could be completely wrong, stoned, and wasted but, didn't MS say they weren't going to optimize Win CE for any processors? Although why they would do that makes no sense at all to me...
I don't fully understand the technicalities of the PPC OS or the ARM chipset so my statement may be compeletely ignorant, but what I don't understand is why we don't see these same issues with Windows NT/2000/XP/etc. If I went from a Pentium II to a Pentium IV, I would see a speed increase with my Windows 98 or Windows 2000 system, despite the fact that Win2000 was not specifically "optimized" for Windows2000. Perhaps this is more analagous to Intel's MMX instruction set or AMD's 3D Now where the software had to take advantage of additional instruction sets?

[Peter Foot]

I think its been stated before that the XScale speed issues is much more an issue of hardware design than OS optimization, its a faster data bus that's needed.

As for whether the next Pocket PC OS will use Wince 4.2 - its already been stated that this version of Wince includes API changes to bring the Pocket PC features into the OS codebase (Pocket PC was a heavily customised version of Windows CE), these forthcoming changes will make future Pocket PC updates run on a more level timetable with Wince OS releases.

Peter

[Andrew]

Tne end of connection mangler as we know it? ;)

[Ekkie Tepsupornchai]

I think its been stated before that the XScale speed issues is much more an issue of hardware design than OS optimization, its a faster data bus that's needed.
I've actually heard both arguments being stated. One saying it was a hardware issue and one saying that the OS needed to be re-engineered to take advantage of X-Scale.

Assuming bus speed is at least partially to blame, my understanding is that bus speed should be transparent to the software, thus all it would take is for Intel and/or HW manufacterers to create PPC hardware with a wider bus, no? Unless the XScale processor is dependent upon the legacy bus.

[FredMurphy]

Microsoft doesn't talk about their Pocket PC OS releases until very shortly before the hardware starts shipping.
I smell an iPAQ with integrated foldable keyboard running WinCE 4.2 with the next Pocket PC OS coming down the road.
That's exactly what I was thinking.

Can't say I'm too worried about whether it's based on CE.NET 4.1 or 4.2. The only imprevement which seems particularly relevant to Pocket PCs would be VoIP for phone models. Having said that, maybe it's exactly why it'll be based on 4.2.

And what they name the new PPC OS really can't be as important as when we get our hands on the damn thing! No surprise that the rumoured Q1 2003 slipped.

The end of connection mangler as we know it? ;)
Announcing..... Connection Mangler.NET™ - mangle your connections faster and smarter. :)

Fred

[Jonathan1]

I'm not waiting around til summer for a new device. I need a new PDA soon so screw it. I'm getting a Zaurus 5600 when it comes out. Supposedly end of January.

It sounds like it will be a year and a half since MS released an update to the PPC. Over 3 years since MS released a REAL update. PPC 2002 is a splash of paint on the system to give it the illusion of an update. The original PPC 2000 was the real deal. The diff between Palm Sized PC's with CE2.11 and Pocket PC 2000 was like night and day.

I may consider getting a PPC 2003 when they eventually ship but for the time being I'm going to use a Zaurus.
:(

[Janak Parekh]

but what I don't understand is why we don't see these same issues with Windows NT/2000/XP/etc.
You have. Intel has been very good at covering up the inefficiences in the processors introduced at each stage of the game by pumping up the clock speed very quickly. You can't do that on handhelds as fast because of the power requirements.

When the Pentium Pro first came out, it actually ran Windows 3.1 slower than comparable Pentiums at the time. The Itanium is currently going through the same growing problems right now, but they've been smart and left it out of the mass market right now, pushing their existing legacy technology as far as it can go.

--janak

[galego2]

I'm not under an NDA and have done research along .NET to learn of Microsoft's plans. Based on what I have been able to find, it appears our Pocket PC 2003 devices (name printed in beta Microsoft material I have seen) should support the following:


1. Bluetooth
2. 802.11 Zero Config
3. Network Media Sensing
4. Real-time Communication
5. Remote Desktop Protocol
6. Internet Explorer 5.5
7. Enhanced Multimedia Support (including DVD, DirectX, and more)
8. Customizable User Interface
9. Multi-screen support
10. Dynamic Screen Rotation
11. Clear Type Support
12. Protected File Storage
13. Integrated RAM/ROM File system
14. Storage Device Drivers
15. Device Notification
16. Security Enhancements
17. Enhanced Language Support
18. XML, SOAP
19. .NET Compact Framework
20. Passport Authentication
21. Ipv6
22. SMS Messaging
23. Office Document Viewers for all Office members
24. Optimized for X-Scale Processors, ARM5

Additionally, in the past the PPC team took Windows CE and then added their own stuff on top of it. The Pocket PC 2000 platform was based on a pre-release version of Windows CE 3 while the Pocket PC 2002 platform was based on the final version of Windows CE 3.

Going forward, Windows CE .NET (4.x) will remain the base for everything and only the user interface will change. Most of the resources that were customizing Windows CE 3 for the PPC have been pulled back. Part of this effort is so the new software can be incorporated faster into devices after being released.

2003 is a big year for Microsoft with a lot of new products coming out in April - July (Windows Server 2003, Exchange Server 2003, Office 11) and Exchange Server 2003 is going to include a lot of enhancements for mobile devices. Additionally, Visual Studio .NET 2003 comes out on April 24 and it will include support for the .NET Compact Framework. This means that at a minimum, we will be able to download the final version of the Compact Framework that day. But, Microsoft will want it embedded in new devices ASAP to improve the customer experience.

So, if I were a betting man, I would predict we will see the new Pocket PC OS sometime in that same time period. Most likely it will be based on Windows CE 4.1 and the 4.2 updates can just be downloaded when they are ready for the devices. But, maybe we will get lucky. And supposedly, Microsoft is promising that the OS will fit in the 32 MB ROM just fine. It will probably come out weighing about 24 - 26 MB.

[Ekkie Tepsupornchai]

but what I don't understand is why we don't see these same issues with Windows NT/2000/XP/etc.
You have. Intel has been very good at covering up the inefficiences in the processors introduced at each stage of the game by pumping up the clock speed very quickly. You can't do that on handhelds as fast because of the power requirements.

When the Pentium Pro first came out, it actually ran Windows 3.1 slower than comparable Pentiums at the time. The Itanium is currently going through the same growing problems right now, but they've been smart and left it out of the mass market right now, pushing their existing legacy technology as far as it can go.
That is true. I do remember that the recent Pentium IV got off to a rather undesirable start and now that I think about it, even the early Pentiums at 60 / 75 / 90 mHz were not convincingly better than a DX4/100. I guess my next question would be then is this performance issue something that Intel must resolve? or is there actually a way to optimize an OS for the hardware. The Pentium Pro as I remember lived a VERY short lifespan and Intel just went ahead and pushed hard with the Pentium II. No this has me wondering if the current XScale could be playing the role of "Pentium Pro" while we await something analagous to the "Pentium II"?

[ExtremeSIMS]

I'm not under an NDA and have done research along .NET to learn of Microsoft's plans. Based on what I have been able to find, it appears our Pocket PC 2003 devices (name printed in beta Microsoft material I have seen) should support the following:

Cool. Thanks, Jaso- er, Galego2.

[seanturner]

Yeah, I believe applications are optimized for certain processors in that I'm pretty sure i've heard intel about trying to get game developers to use their compiler...

Not to mention a lot of debate in the SETI@home community about wanting specific processor optimizations....

[Ekkie Tepsupornchai]

Yeah, I believe applications are optimized for certain processors in that I'm pretty sure i've heard intel about trying to get game developers to use their compiler...

Not to mention a lot of debate in the SETI@home community about wanting specific processor optimizations....
If XScale is the only game in town, I don't understand what difficulties Intel would get in convincing game developers to use their compiler...

[seanturner]

Its not really the only game in town though. It is really only included on the last two revisions of devices and many people still have legacy devices.

But, yeah, I have no idea why I can't name a single X-Scale optimized app....

[Ekkie Tepsupornchai]

Its not really the only game in town though. It is really only included on the last two revisions of devices and many people still have legacy devices.
That's true... though those numbers diminish each day... MIPS and SH3 are dead in my book, so the only other processor developers would need to worry about is the original ARM. My guess is that a kick-a** games optimized to XScale would be a welcome site to everyone, I mean it's a sign to all that the faster processors really are worth something! If I were Intel and MS, I'd pay for that development.
But, yeah, I have no idea why I can't name a single X-Scale optimized app....
Actually, isn't there an XScale version of PocketTV? That's the only one I can think of though.

[seanturner]

Actually, isn't there an XScale version of PocketTV? That's the only one I can think of though.

There might be... I'd kill for an X-Scale version of divx though...

[seanturner]

&lt;oops... duplicate post>

Yeah, that's happened to me three times in the last 2 days.

Is there something going on with the server?

And i've also noticed that my preference to always attach my signature has been somewhat flaky; sometimes it does it, sometimes it doesn't have the box checked...

[Ed Hansberry]

I know it is being picky, but there are no X-Scale optimizations.

The optimization would be for ARMv5 chips, of which the X-Scale is one of, as are most (all?) of the ARM processors PalmOS has settled on, like the newer TI-OMAPS. The StrongARM is ARMv4. ARMv5 chips can run ARMv4 code, but they do a bit of emulation/translation to do so and that is why a 400MHz X-Scale runs about the same as a 206MHz StrongARM for identical code.

The downside is, any ARMv5 optimized code won't run on a StrongARM or any other ARMv4 chips, so as this trend continues, iPAQ 3600/3700s, Jornada 56x, Toshiba 310, 570, etc. will begin to get left out.

[Janak Parekh]

I guess my next question would be then is this performance issue something that Intel must resolve? or is there actually a way to optimize an OS for the hardware.
Of course there are ways to optimize OS's - compilers target specific new hardware versions. For example, Windows 2000/XP need at least a Pentium; 386's and 486's won't work anymore. However, as I just noted, these shifts tend to be very, very rare. The fact that Windows XP still supports Pentium, which is now several generations old, is telling.

There are other problems, as well. Faster clock rates mean deeper pipelines. Deeper pipelines introduce all sorts of problems with complex machine instruction mixes. It's a very, very hard thing to do to keep on pumping up clock rates.

The Pentium Pro as I remember lived a VERY short lifespan and Intel just went ahead and pushed hard with the Pentium II.
Not that short. The Pentium Pro was popular in the server market, as it performed pretty well with NT, which had mostly 32-bit code. It just sucked really badly at 16-bit code. (Transitioning to 32-bit was a much, much bigger thing than just XScale optimization, mind you.)

The Pentium II didn't fare much better, but we had nearly finished the transition to mostly 32-bit code by that point, and Intel boosted the clock rate to cover up the PPro's performance problems with older OS's.

No this has me wondering if the current XScale could be playing the role of "Pentium Pro" while we await something analagous to the "Pentium II"?
Well, from the other issues I've heard (slow emulation of ARMv4, memory bus issues), I guess at a high level, sure. :)

--janak

[Janak Parekh]

ARMv5 chips can run ARMv4 code, but they do a bit of emulation/translation to do so and that is why a 400MHz X-Scale runs about the same as a 206MHz StrongARM for identical code.
It's not clear, though, that ARMv4 emulation is slow on every implementation. It might be on the XScales, but on the newly announced Samsungs we might see much better performance. The question is if we should be blaming Intel for the poor emulation performance or Microsoft for leaving people out in the dust. I've read reports that ARMv5 optimization, at least on Linux, isn't going to do anything like double the XScale's speed, but I don't know where to look for these right now.

--janak

[Ed Hansberry]

It's not clear, though, that ARMv4 emulation is slow on every implementation. It might be on the XScales, but on the newly announced Samsungs we might see much better performance. The question is if we should be blaming Intel for the poor emulation performance or Microsoft for leaving people out in the dust. I've read reports that ARMv5 optimization, at least on Linux, isn't going to do anything like double the XScale's speed, but I don't know where to look for these right now.
Yup. Lots of fingerpointing between Intel and MS on this. As to the Samsung - there are other issues. I think the bus of the X-Scale is the same width as the StrongARM - not good for a double MHz chip. The Samsung may have a fatter bus which will help matters.

[seanturner]

Yup. Lots of fingerpointing between Intel and MS on this.

Actually, has either Intel or MS said anything about these performance issues?

[Ed Hansberry]

Actually, has either Intel or MS said anything about these performance issues?
Lots back in the summer when the X-Scale PPCs came out. Our archives should have links to cnet and other articles.

[Ekkie Tepsupornchai]

I know it is being picky, but there are no X-Scale optimizations.

The optimization would be for ARMv5 chips, of which the X-Scale is one of, as are most (all?) of the ARM processors PalmOS has settled on, like the newer TI-OMAPS. The StrongARM is ARMv4. ARMv5 chips can run ARMv4 code, but they do a bit of emulation/translation to do so and that is why a 400MHz X-Scale runs about the same as a 206MHz StrongARM for identical code.

The downside is, any ARMv5 optimized code won't run on a StrongARM or any other ARMv4 chips, so as this trend continues, iPAQ 3600/3700s, Jornada 56x, Toshiba 310, 570, etc. will begin to get left out.
Actually, I don't think this explanation is picky at all. The "optimization" misnomer makes a lot of sense now. It's not that we need an OS optimized for XScale, it's just that we eventually need an OS that doesn't run in emulation mode to accomodate the ARMv4.

[Ekkie Tepsupornchai]

The Pentium Pro was popular in the server market, as it performed pretty well with NT, which had mostly 32-bit code. It just sucked really badly at 16-bit code. (Transitioning to 32-bit was a much, much bigger thing than just XScale optimization, mind you.)
Didn't know the Pentium Pro was hitting big with the servers... but I was in my final year at college when it came out and I was more concerned with building my own consumer desktops at that time...

The Pentium II didn't fare much better, but we had nearly finished the transition to mostly 32-bit code by that point, and Intel boosted the clock rate to cover up the PPro's performance problems with older OS's.
Makes perfect sense. Thanks.

[FredMurphy]

Yet another thread gets tangled up in the xScale debate. :roll:

[Ekkie Tepsupornchai]

Yet another thread gets tangled up in the xScale debate. :roll:
Yeah, it's like bad fungus... it never goes away... :?

Guess it shows how twisted this whole XScale thing is...

[Pony99CA]

Needed features:

1) X-Scale optimization
2) Support for up to 1Gig of native RAM. PocketPC
3) Support for up to XGA resolution devices.
4) OS driven support for redefining the screen orientation.

5) OS support for Firewire and USB On-the-Go

6) Native Bluetooth and Wi-Fi support.

7) Native DirectX for PocketPC. A Comprehensive Graphics and multi-media libarary so that vendors will include 3D hardware.

8) USB hosting for connectivity with mass storage devices.
I nominate Will T. Smith for president....no ruler for LIFE, of the PDA universe.
And like a politician, he's promising more than he can deliver. :-) I doubt we'll see support for 1 GB RAM -- we're still at 64 MB (although some after-market vendors can cram up to 256 MB in). XGA? Come on, VGA would be a major step up. Firewire? Read the home page article for why that won't happen.

Steve

[Ed Hansberry]

And like a politician, he's promising more than he can deliver. :-) I doubt we'll see support for 1 GB RAM -- we're still at 64 MB (although some after-market vendors can cram up to 256 MB in).
That is hardware. CE 3.0 supports 512MB now.

[Ekkie Tepsupornchai]

XGA? Come on, VGA would be a major step up.
Assuming we're looking to maintain the current size of these devices, I can't imagine how XGA would be beneficial enough over VGA to justify the cost that would surely be associated.

Seriously, what exactly do we need from XGA resolutions in a 3.5" screen that VGA couldn't adequately provide?

We could enlarge the screen of course, but to really make the most of XGA, we'd be approaching something the size of sub-notebooks.

[Pony99CA]

Assuming we're looking to maintain the current size of these devices, I can't imagine how XGA would be beneficial enough over VGA to justify the cost that would surely be associated.

Seriously, what exactly do we need from XGA resolutions in a 3.5" screen that VGA couldn't adequately provide?

We could enlarge the screen of course, but to really make the most of XGA, we'd be approaching something the size of sub-notebooks.
Well, most iPAQs have a 3.8" screen, and the Toshiba e550G has a 4" screen, but I still wonder how good XGA could be on those.

My Sharp Mobilon Handheld PC only supported 640 x 240 graphics, and its screen was 6.5".

Steve

[FredMurphy]

More evidence of when it will be released and what it will be called from Microsoft Press (http://www.microsoft.com/mspress/books/companion/5960.asp)

Visual Studio .NET 2003 includes emulators of a Windows CE .NET and a Pocket PC 2003 device....

Visual Studio.NET 2003 is due for release 24th April, and surely the real OS can't be far behind the emulated version!

Fred