MS Tablet PC: 70 per cent hype, 30 per cent snake oil
And somewhat beneficial to Redmond's profit margins, we expect...
Bill Gates got another chance to wave Tablet PC prototypes about at his Comdex keynote last night, and he got to announce some more partner companies for the project too. But isn't it time we asked a few questions about the Tablet PC? Such as, what is it, what's so special about it, and why is it taking so long?
The answers to these questions seem self-evident until you set out to look for them. As Bill keeps telling us, it's amazing, lightweight, long battery life, you'll be able to carry it around with you all day, and it'll be "the most popular form of PC within five years." But although the Tablet PC has its own little section  on the Microsoft web site, and the morlocks are due to post snaps in the Microsoft Comdex virtual press office later today, there's practically nothing in the way of hard information.
There is no published specification, no hardware design guide, very little at all apart from whatever it was Bill said about Tablets last. But here, for what what it's worth, are the basic facts that can be nailed down. The Tablet will run Windows XP Tablet Edition, which will have all of the features of Windows XP Professional, plus some Tablet-specific ones. So the next version of XP could have been called XP Pro Plus, if somebody else hadn't got there first. Gates did announce the first Tablet PC SDK yesterday, and there's a walk-through of this available here,  but the specific focus here for developers is on how you deal with pen and voice input. Fans of recent trends in Microsoft user interfaces will no doubt be dismayed to meet the green parrot in the voice section; you have been warned.
Pen and speech input and recognition that worked in a handheld device certainly could be seen as revolutionary, and certainly could explain the long run-up for the machines, which aren't due on the market until the second half of next year. But with the SDK only out now, and in beta form at that, software developers aren't getting much of a window to design anything really earth shattering. But Microsoft itself has been busy for some time, and quite a lot of the groundwork for Tablet has already been done in Office XP and Windows XP.
It's also worth noting that the Office XP support announced yesterday will include the ability to integrate handwritten notes in documents and send handwritten emails via Outlook 2002. From the point of view of form factor that's key, because quite a few of the Tablet variants will have detachable keyboards, and users will need to be able to have some functionality for their apps via handwriting and voice input if they're going to be just carrying the screen unit around. So we can presume that most of the first wave Tablets will have Office XP bundled, and that Microsoft will have got closer to its goal of selling a copy of Office as well as an OS with every PC that ships. Corel is also one of the partners, so there will be at least one rival here - but how strong, how effective?
Handwriting and voice also have some impact on the shippability of the Tablet from a hardware perspective. The Microsoft line is that it's only just becoming possible to do Tablets because the hardware is getting powerful enough to support these, but actually we're well into 'how long is a piece of string' territory here. Hardware requirements will depend on the amount of crunching the software has to do, and that itself depends on how much Microsoft decides it wants the software to do, and when. One might consider that the ETA has more to do with Microsoft's product development and financial requirements than with today's hardware being incapable of doing the job.
As regards the hardware spec itself, at the moment we have to infer this from the public statements Microsoft has already made, and from the nature of the partners companies. The original chip partner was Transmeta, but Intel climbed aboard just in time for the first announcement. Via has now joined in too. Whichever company's chips a specific implementation runs, they'll have to be low power. There will be no fans, the units will have to run cool enough for it to be comfortable to carry them around (think how people would have laughed if you'd predicted portables too hot to carry ten years ago). They'll also need sufficient battery life for you to be able to carry them around without docking all day, and they'll need to be able to resume in under three seconds.
Temperature and battery life requirements make it obvious why Transmeta was in there early, but the laxity of a related requirement indicates that Microsoft isn't demanding things that portables can't do already. Tablet PCs should be able to survive in suspend mode, i.e. power still going to memory, for three hours. This is not onerous by the standards of current notebooks, but does beg questions about what kind of Tablet PC will be able to last the whole day, and how much of that time it's going to have to spend in hibernate mode in order to do so. It does however seem perfectly possible that the base hardware for a Tablet won't be wildly different from, say, a Transmeta-based Vaio.
Tablets will come in all shapes and sizes, Microsoft tells us, but will be about half the weight of today's notebooks. Some weight can be lost if you minus the keyboard and the heat-related items, and as they will be "legacy-free" there may be some related savings there. A CD drive is an obvious saving, given that one of the ideas associated with Tablets is that you truck them around under your arm and plug into docking stations. They're supposed to be your main, personal computer. Spec details for the prototypes revealed yesterday, from Acer, Fujitsu, Compaq, Tatung, Viewsonic, Toshiba and FIC, are sparse, but Acer intends to use a 10in LCD, so size will also affect weight.
And it clearly has an impact on something else too - build cost. A pen-enabled screen is clearly going to be more expensive,* but a 10in screen offers considerable savings over the sizes typical in notebooks these days. His Billness also says that they'll cost around the same as a mid-range notebook today, so even before you consider that some Tablets' form factor will require a screen smaller than 10in, you can see money to play with, and why manufacturers might be interested.
But what they get out of it will be crumbs, transitional ones at that, compared to Microsoft's reward. Prior to coming up with the Tablet concept Microsoft parted company with Intel, which had previously worked with it to define hardware standards for the PC market. The last ever joint effort, the PC 2001 spec, in fact provides the baseline hardware spec for Tablet, but the specific details are now being defined by Microsoft, and Microsoft alone.
Microsoft gets to say who is and who is not allowed to the Tablet PC feast; this gives it more power, and is also helpful in that it means Microsoft can put the focus on a narrow range of hardware and components. This (similar to the Xbox approach) makes it simpler to produce reliable software, and of course gives it even more power. Microsoft also gets the revenue from the OS and (in many cases) the Office licence, and The Register wouldn't be in the slightest surprised if there turned out to be other, design and hardware-related licence fees in there. There will no doubt be other rake-offs involved - what, for example, is Zinio Systems, purveyor of "digital magazine technology and services" doing in the Tablet supporters list?
All of this, remember, is associated with a class of product that Bill Gates intends to be "the most popular form of PC within five years." Microsoft's control of this form of PC will be far greater, and the revenue Microsoft derives from it will be far greater. Come the second half of next year it will be hyped to the skies, and if Microsoft has its way, it will take over the business. But what is it? It's a notebook computer with a detachable keyboard and voice and pen input.
We can do these already, people have tried them already, and if this one succeeds it will be the ultimate triumph of packaging over substance. ®
* Register digression footnoted in the interest of order: The author's local rail station has recently re-equipped with embedded Windows ticket machines. These slow the journey by 5-10 minutes because the ticket machines have consequently become prone to BSODs. The UI is clearly touchscreen, but it is presumably cheaper to not use a touchscreen and use arrowed buttons that point at the bits of the screen you'd touch if it were a touchscreen. Thus, queues are slowed significantly by people unfamiliar with these particular machines who intuitively jab the screen, rather than push the buttons, because it's obviously a touchscreen. Factor in another 5-10 minutes delay per day.
MS XP-Tablet combo cuts Intel out of PC standards biz