Microsoft, what we need is Vista Mobile

Vista doesn't work well on UMPCs and other mobile PCs as they are shipping today.  I'm not the only one who's been ranting on this lately, you don't have to look far to find consumers who are not happy with the current situation in the mobile PC space.  I have stated (and prophetically some would say) that Apple is the only company who can turn out a mobile PC with the requirements that consumers need.  To be fair I also stated that the other company that could do this is Microsoft.  This article is a look at how Microsoft could make an operating system/ mobile computing environment that would revolutionize this space and also leave Apple in its dust.

Operating system environment

Microsoft has thrown everything into Vista and its incorporation into all PCs, including mobile PCs.  This is a huge mistake because the OS environment is too big, too resource hungry and too hard on battery consumption to ever work well as it's currently shipping.  What is desperately needed is a new OS platform for mobile PCs, I'll call it Vista Mobile for this article.

Vista Mobile could be produced to provide the operating environment that mobile PC users need and demand.  It should be designed to provide compatibility with bigger PCs in the user's world, utilize power optimally, and this should be done in such a way to still allow users the ability to run the programs they need to run to be productive.  So how can Vista Mobile be produced to maximize the user's experience based on this criteria?  I'm no hardware/ software engineer but I am pretty familiar with the way these things are done today so here are my thoughts on how Vista Mobile should be produced.

Basic environment

I firmly believe that for mobile devices to be the best they can be that the operating system needs to be embedded in the device and not installed on spinning media.  This goes a long way to addressing all the criteria I've listed and more importantly gives a way to have the mobile PC work with instant on and off.  This is very important for the mobile user to be able to hit the button and have the device instantly return to where it was when the user powered it off to save battery.  This is a no compromise area of design because the positive impact on the user experience cannot be greater.  I know that Microsoft is looking to move entirely to the Vista platform but if Vista is not the best to provide an embedded environment then it shouldn't be used here.  Embedded XP is being used all over the world today and if it would make for a better starting point for Vista Mobile then so be it.  Consumers will not complain, believe me.

Having the OS embedded in the hardware also can be capitalized to make the overall performance of mobile PCs to be stellar.  The current model of having the OS live on a spinning hard drive and constantly pulled into memory as needed is not optimum for mobile PCs and having it always sitting in memory could only be a performance plus.  The embedded portion of the OS can be designed to coexist with existing processor technology which makes sense given how much work Intel has done in this area.  The McCaslin processors and upcoming Menlow processors don't handle Vista well but if Vista Mobile was designed with these lower power processors in mind we could have some pretty killer performance with them.  We don't need to reinvent the wheel and plan on new hardware, let's use what we already know is out there and upcoming.  Vista Mobile could be optimized for these processors and costs would be contained and performance enhanced.  I can envision a pretty killer mobile PC under these circumstances.  Mobile PCs should be faster than their bigger counterparts and should be one of the objectives of producing Vista Mobile.  I am convinced it can be done.

One of the areas that conventional operating systems fail miserably for mobile PCs is because they are designed to handle as many different hardware components as possible.  They have to do this because there are no real standards in the hardware world, graphics chipmakers are always updating their products for example.  Vista Mobile doesn't need this open-ended component support.  The opposite is better in fact, there is nothing wrong with Microsoft cherry-picking the best fit graphics support, Bluetooth support, WiFi support, and any other areas we need in mobile PCs.  Pick a few and build bullet proof support for them in the embedded OS.  That just makes sense and will insure the user has no problems with the hardware.  Vista Mobile could even be optimized to get the best performance and battery life with them.  That can only be a good thing with Vista Mobile.

Vista Mobile should be written to allow users to add the programs they need to get their work done.  Forget spinning media, though, because it's slow and hard on battery life.  No, write Vista Mobile to use solid state disks (SSD) from the ground up.  While SSDs are beginning to take off the ones big enough to run conventional Vista are too expensive and frankly with the OS embedded in memory the larger capacity SSDs are not needed.  Vista Mobile should be written to require an SSD for user storage and application installation and that could easily fit in a 32 GB SSD.  Costs would end up the same as spinning media as mobile PCs began to fly off store shelves and the performance benefits cannot be overstated.  No moving parts are what we need Vista Mobile to require.  While you are at it Microsoft one thing you can do that would have a huge impact on our mobile PC performance in our mobile work life is to embed Microsoft Office.  Yep, you heard me right.  MS Office is a huge software suite that has a negative impact on storage requirements and operationally for mobile PCs.  Embed that right into Vista Mobile and you would turn a negative into a huge positive impact on performance and battery consumption.

Vista Mobile should be optimized for both touch and pen input.  Mobile PCs provide a more productive experience when they can be manipulated with the fingers, but pen input is critical for data entry when needed.  This could easily be implemented right into the OS using the existing technology for both touch and ink in Vista, and palm rejection technology could be built right into Vista Mobile.  There are not that many hardware digitizers being used in mobile PCs today and Vista Mobile could be tweaked to provide the best inking experience with all of them without a ton of work.  The goal with Vista Mobile is not to compromise but to optimize.  If it's needed to produce a very constrictive design reference for screen sized supported then so be it.  I'd rather have a great mobile device with fewer screen size choices than poor ones any day.  We shouldn't view that as restrictive rather as a way to get an optimum user experience.  I am all for that myself.

Vista Mobile should support interfacing mobile PCs with the desktop environment in its core.  Mobile PCs should have a dock that has the device instantly connect to the user's computing environment without problems.  The ability to auto-sync users documents with a desktop would be great but at the very least the SSD on the mobile PC should show up as an external disk drive on the desktop.  If the user has to manually copy stuff over so be it, it's better than we have now.  Mobile PCs will be used as primary computers by some but for most they will be companion devices and this type of PC interaction is not a difficult thing to implement.  Dock your mobile PC and it becomes a part of your desktop environment, I think everyone can imagine how great that will be.  Docking should also allow tethering the mobile PC to peripherals for those who will use them as a primary device.  Imagine how great this scenario could be- you dock your mobile PC, the desktop sees it and mounts the drive, Vista Mobile detects the presence of another computer and asks you if you want to use its screen, keyboard and mouse with the mobile.  The technology exists to already do this today so lets put it in Vista Mobile.  That would be drop-dead wonderful.

This is by no means a detailed technical how-to for Microsoft to produce Vista Mobile, like I said I'm no engineer but there are hundreds (if not thousands) of them in Redmond who could do this.  Mobile PCs are not used the same as other PCs and will never work optimally using operating systems designed for desktops.  We need Vista Mobile and more importantly so does Microsoft to move forward from the debacle of Vista.  Let's all let Microsoft know how badly we need Vista Mobile as the computing space gravitates to smaller devices that are designed to be good work machines and lifestyle devices concurrently.  We need your help, Microsoft, and you can do this if you will.

Comments

[Mickey Segal]

Having a crippled version of Vista is not a good idea. One of the advantages Vista tablets have over the iPhone line is the ability to be docked and act as a full computer. What Microsoft should do is make it easy for the user to disable certain functionality while mobile and Microsoft should fix the problems with docking / un-docking. Adding the ability to support real telephone capability and have such capability always on would also be a plus.

[James Kendrick]

Mickey, I described a totally compatible OS optimized for mobile devices.

[Aaron J. Walker]

Great article James (although I still disagree with the first one in your series), I think you hit the mark with talking about Vista Mobile. It continues to boggle my mind why Microsoft does NOT embed the OS (whether XP or Vista) in all mobile devices, including laptops. You would think they would have learned how-to and all the benefits of from Windows Mobile. It just makes sense to me. Leave all the fancy stuff to the desktops.

Windows Mobile is embedded, has instant on, long battery life and (nearly) seamless integration with the desktop. Take the concept and apply it to all products (or as many as feasible) in the Microsoft line for mobile devices and laptops. It's like one hand doesn't see learning from the other hand. That's just plain dumb.

Updates could be handled by flashing the ROM. Consumers could get the choice of which programs to have embedded or not. ("You'd like Office 2007? Would you like us to pre-instal it to ROM?")

Seems a no brainer to me.

[Sumocat]

James, have you been reading my rousing debate with MiniMage regarding Microsoft's flawed approach with Windows for mobile PCs?
http://www.gottabemobile.com/CommentView,guid,0dc9687e-169c-4a4f-ba71-359521a6fbdf.aspx#commentstart

The gist of my argument is Apple streamlined a version of OS X for the iPhone with a tailored interface (and per MiniMage, it can still run Unix desktop apps), while Microsoft actually made a heavier version of XP (and kept that weight in Vista) for mobile PCs with the same desktop interface. To me, that's a major detriment to the mobile PC market, and a big reason why many are looking to Apple to bring the form factor to the mainstream. Five SKUs for Vista, not one designed for mobile computing. Tell me how that demonstrates support for the form factor.

[Scotty]

Vista Mobile doesn't begin with new requirements, it starts with new development processes and team spirit for the MIcrosoft Developers. The limit on Microsoft's ability to deliver quality new OS' isn't their vision: its their ability to write them.

For another example just look at Surface. Every body else is doing 16:9 HD but Surface is 4:3 1024x768?!!? They've been working on it since 2001? And what can it do? Rippling water, crude paint program... where is the NY Times? Where are the dazzling rich color high definition coffee table books and photo albums? Six years of development and it can't show a newspaper on demand?

Sorry James, they can't handle writing Vista Mobile.

[Tyler]

James,
While I think that you have an excellent point, I think that you may be placing the blame on the wrong company (to some extent). As you have mentioned, MS has developed XP Embedded, which, virtually is everthing you ask for (except based on XP rather than Vista). So, lets say (hope?) that MS releases a comparable OS -- Vista Embedded (seems to be an adequate assumption since they did XP, and NT 4). While that would answer your request, there is still one essential element that is missing - the device manufacturer. No matter what MS does, until they decide that they will release a HDW device, they (and therefore, we) are constrained by the folks who put out the devices. Have you ever seen a Consumer Focused UMPC/MID type device on XP Embedded (I think that the only one I remember is the Smart Displays, and those were tainted with short sightedness b/c of Marketing and Hardware limitations)? And that has been around for years!
I think that what we should be asking for is a device manufacturer to step up and create the device using XPe, or help to pressure MS to Create VistaE. Based on the literature for XPe, it can be installed in less than a GB, you can choose which elements to install (limiting device drivers to only those that are needed), and its instant on/flash based, etc. Seems like a winner, if we had the device to run it.....

[D]

1. embedded OS does NOT mean instant-on, i have no clue how you guys got this idea

2. Windows Mobile is NOT instant-on, it's just sitting in stand-by. if you want to see a WM device boot just do a soft reset, it typically takes 20-30 seconds to boot.

[Chuck]

James - I agree with most of what you say (I can live with spinning media, especially for secondary storage), but you left out the one thing that will remove the most bloat...

How many users does your Mobile PC have? Why does it need to be multi-user? The amount of code required to support multiple users on a PERSONAL computer is significant, to say nothing of it's uselessness on a Mobile PC.

[Nikooo]

One can push the thinking a little further...

What about creating a ROM flasher that would allow the end-user to select what software should be implemented in the ROM in addition to the tailored system?
Would be cool to put MS Office in the DVD drive, run the flasher... and boom it is added to the ROM :)

[Aaron J. Walker]

Tyler makes a good point.

I've *heard* about XP embedded, but I've never seen it or been offered it as an option on any pc, UMPC or Tablet PC. Microsoft would need to push to make that an option to the consumer from their partners.

Perhaps there is no Vista embedded because XPe (isn't) didn't go well.

I would take VistaE, XPe as well as an OfficeE on a UMPC or Tablet in a heart beat for the speed and battery life.

(agreeing with Chuck)And yes, kill all the multi user profiles on mobile devices. No one else uses my UMPC or Tablet unless I hand it to them. Desktops, sure, they are often shared in homes and such but a UMPC? Not often enough where there would need to be seperate user profiles.

[Jake]

I would love a Windows device with instant-on capabilities but I think what you describe James would end up as more of a hinderance. Just look at the trouble people used to have with Windows Mobile when a new version was released. Pocket PC manufacturers would be slow to release the new software or would refuse to altogether. Having the OS tied to the hardware to such a great extent is never a good idea - just take a look at Apple's market share.

I think you just need to wait for SSDs to come down in price and that should give you the speed you're longing for without having to make other sacrifices. I know I can't wait to be able to stick one in my P1610 but I suspect that we are a few years away from them being generally available in the sorts of capacities that people actually want (32Gb just doesn't cut it anymore). Heck, I can't even find someone selling these things on the internet yet.

[Jake]

@Chuck, this is the problem with a cut down version of Windows. Although you may not think it's necessary to support multiple users on a mobile device others might. I don't think parental controls are necessary in Windows at all but other people do. Where do you draw the line?

[James Kendrick]

I'm not placing the blame on anyone with this, just stating what I think needs to happen to make a mainstream mobile PC. You're all right, where do you draw the line as to what's included and what not? For me I think keeping it as simple as possible is more important than extra features. It's the "full experience" OS that gives us all the problems with mobile PCs today.

Give me something streamlined and light and I will be happy but more importantly so will Aunt Sally and Uncle Bill.

[GoodThings2Life]

I've always preferred the Windows 95/98 approach to setup... allowing a custom install where users can, during the initial OS install pick and choose the individual components that they want.

I also **HATE** having what-- 5 or 6 "flavors" of Vista, i.e. Home Basic, Home Premium, Business, Ultimate, etc... although I understand their logic for the madness.

As a result, it seems to me that the best option for Microsoft is to mix and match the best aspects of those options as follows:

During setup have a "Machine Profile" selection. If you choose "Multimedia Desktop" it will preconfigure all the Media Center Extensions. If you choose "Tablet PC / UMPC" it will preconfigure the system with a lightweight mobile experience with inking functionality. So-on-and-so-forth.

Also, on that screen it would have a "Custom" checkbox to allow the user to pick a predefined profile and then customize it further.

[bluemonq]

If MS started talking about "approved" graphics chipsets and like hardware, I dare say we'd start hearing complaints about MS's abuse of its monopoly powers and anti-competitive practices.

[Jake]

Also, don't forget that Microsoft went down this approved hardware route with XP Media Center. As a result, it wasn't available off the self and hardly anyone bought it - despite the fact that it's actually a very good product.

James, have you got a list of features you would be prepared to lose in a streamlined version of Windows? To be honest, the only way I think you could please everyone would be to make adding/removing features a bit more straight forward.

One other thing that occurred to me is that if you left an embedded version of Windows up to the OEMs you'd probably just end up with an almighty amount of crapware that you couldn't get rid of.

[Chris K]

Services, services, services!

The thing that kills you on Vista, more than anything else, is all of the heavy lifting, and believe me, there's a lot of it.

Vista is hell-bent on providing you with access to ALL of your data, ALL of the time, as fast as possible by the time that you request it.

This means that it caches aggressively, moreso than any other OS that I've ever worked with. This is great, if you've got at least 2GB of RAM and a nice, speedy disk. Almost any modern CPU is overkill.

Of course, at the same time, it's obssessed with giving you the shiniest chrome imaginable, which often involves a dedicated GPU with its own RAM, or overtaxing an embedded GPU on shared RAM.

Vista needs to ship with a Mobile Optimization Wizard, as this would solve most of the problems. Here's what it needs to do:

1) Run an iometer-style test on the existing disk, testing a series of random and sequential accesses at various ratios, on both reads and writes.
2) Keep a log of battery life.
3) Ask the user a few questions. Do you want multiple user accounts? Do you use your computer all day long, etc. Log how often the user puts the system to sleep each day.

Use this data to determine what can and can't be done. Disable Fast User Switching, lock down filesharing services, download ONLY security patches, and keep those small. Disable caching services, lighten up on the Aero interface, etc.

On the desktop, you can afford to optimize when the user's away. On the laptop, the user's rarely away. On a UMPC/micro PC/ultraportable tablet, the user is best served by having an OS that assumes that if it's on, it's going to need to run full-tilt before going right back to sleep. There's no reason that Vista can't be trained to do this, but how many users are really willing to delve that far into disabling services, tweaking the registry, and killing off 99.9% of the "Wow?" Microsoft should just admit that sometimes, the "Wow" factor just gets in the way of getting things done, and set up a massive monolithic wizard to cut it down.

[Ayrkain]

Having done massive amounts of optimization on Vista - disabling every service I could afford to, giving up all the eye-candy, and most of the "Wow"(?!) - I've given up. I have a UX180P with a 32 GB SSD drive and it still wasn't enough to make Vista a good experience. I brought it up to the point that it was almost as fast as XP on the standard drive. I downgraded yesterday, and the speed difference is incredible. I still run 2 copies of Vista on my tablet PC and my desktop, but Vista is just not Mobile PC friendly.

[Rodfather]

I don't think Windows is a good idea at all for mobile devices. It's too bloated and one of the reasons is to be able to run all the legacy apps going back to 95/98.
Better off being as efficient as possible but still have something exciting for developers.
Linux + Adobe AIR/Silverlight?
Intel thinks so.

[Opus]

Not one more Windows Vista edition, if they do call it Vista Dumber or Windows VD.

Fix the OS, PC makers fix your drivers, and cpu/gpu makers improve your chips for both performance & battery life, let's not castrate everything to make it all work!

I can't believe what I'm hearing...

[Alan]

Since the point is to always be mobile, the system needs to be the main one, and not sync'ed or slaved to another. Along with portability come the functions of a true PIM, so why not provide for Outlook's calendar to wake the hardware from standby? Providing the user with audible reminders of appointments is very basic and should not require that the system be fully powered on and logged in.