Guy Kewney

The future of the notebook PC depends on whether people find the Tablet PC an acceptably accurate handwriting transcriber.

If it can genuinely make hand written notes and search them with a text engine afterwards, then there's a base on which to answer the question: 'How powerful does a notebook PC have to be?'

Very roughly, a Tablet PC has to have the processing power of a 1GHz Pentium III processor, about 256MB of internal memory and a hard disk. But a notebook PC doesn't have to do anything of the sort.

The screen is defined by the internet - it has to be 800 by 600 as an absolute minimum and full colour. It needs to be able to store a day's email messages, all the names and addresses I'm likely to want to refer to, and any voice recordings I want to take.

It also needs a proper keyboard and be capable of running an operating system like Windows.

OK, now have a look at the Psion Teklogics Netbook Pro. It's got a proper keyboard, and a nice, prop-up full colour display. It runs a version of Windows, more or less, as well as Outlook and Internet Explorer.

The thing is, it runs on a 400MHz XScale processor - yet another variant of the ARM. Yes, it will surf the web and edit Word documents.

It probably won't do proper Excel spreadsheets, but it will take all your Outlook data, contacts, email and appointments.

Although it doesn't have or need a disk, it does have internal Ram and Flash, and some Rom, and that's it. What it can't do, though, is Tablet-level handwriting recognition. That takes more storage.

But does it have to? As I said at the start, we don't know. Handwriting recognition is now as good as it's going to get.

Next year, IBM will launch its first Tablet PC. If it succeeds, then we know that a notebook PC will have to be able to run Windows XP Tablet edition, and the Notebook Pro and the Dana Palm powered notebook PC alternative (from AlphaSmart) aren't going to be the ancestors of a new generation of low-power, compact workbooks, all based on ARM variants.

If it flops, then I'd predict a sudden flood of such Windows CE and Palm and Symbian personal computers. Because in five years' time, the amount of power you'd actually need to carry around with you will be a rendering engine, a voice digitiser, a display chip set and a communications controller. Everything else can go remote.

You will be connected at 512Kbps to the internet whenever you need to be, and that's enough to let you work on the screen of a remote display using RDP, store data on a remote storage system, and use the remote processor for heavy tasks like text analysis and voice recognition.

It took IBM 11 years before it launched its first ThinkPad in 1992; it took eight years to sell the first 10 million notebook PCs, but only three to sell the second.

As to how many they'll sell by 2014, it depends on whether people still buy desktop PCs. If they don't, then ThinkPads and other notebooks will replace half the desktop business. After all, why would you buy something big and noisy when you can buy something smaller with a built-in UPS for when the power blinks?

But if desktops remain popular, then I would guess that there's a good chance that the next 11 years won't see more than 30 million ThinkPads sold, unless IBM builds a Windows CE, or Symbian, or Linux or Palm-based compact ThinkPad using an ARM chip.

I think they'd sell about 300 million of those, unless we have to have something powerful enough to run a Windows Tablet Edition.

Like this story? Spread the news by clicking below:

 del.icio.us     Digg this     reddit!