Microsoft Office 2003: It matter$

There was, for once, some substance in the hype surrounding the launch of Microsoft Office 2003, although the marketers could not resist going over the top with a breathy video proclaiming that it will turn us all into Shakespeare.

The new version is a step change almost as big as the one that brought the world to your desktop with links to the web, because the extensive support for XML turns Office into a potential front end for any database, online service or trading system.

That might not mean much to the average end user, but it is a major shift for developers of next-generation systems. There was, nevertheless, a feeling at the launch that Microsoft may at last be reaching the limits of its dominance.

Some of the new features for collaborative working are similar to what Lotus has been doing for years, and even XML is a weapon that can backfire. It is an open standard, and the more Microsoft promotes its use, the more opportunity there will be for rivals.

Microsoft will try as usual to lock users into proprietary 'extensions', helped by XML, which allows for a limitless number of dialects, and it will succeed to a degree.

It is, however, operating in a very different world from the one it cornered two decades ago. The days have gone when people would queue at midnight to get new versions of software, and wish lists have given way to complaints about feature overload.

Microsoft cannot earn money by adding new features for ever. It tried to head off this crisis by moving from selling software as a product to leasing it as a service.

But this has had the effect of focusing minds on the fact that governments and other large organisations are committed to paying Microsoft tens and even hundreds of millions a year for years. Naturally they are looking at alternatives.

Intel founder Andy Grove is not alone in questioning how much longer Microsoft and other US software firms can maintain the revenue and profit levels of the past few years.

Only in the software industry, with its monopoly-friendly need for a standard platform, could prices have been sustained as long as they have in face of massive global sales. Other countries are perfectly capable of producing good software and are starting to do so.

I put these points to Steve Sinofsky, senior vice president of Microsoft's Office project. He explained that the new Office is just the start of a project to "increase the productivity of knowledge workers".

"The idea that we have done all we can [to improve] the synthesis, analysis and reporting of information ... I just don't believe that for a second," he said.

Sinofsky admitted that most people use Office only for relatively simple tasks, but he cited a recent poll purporting to show that as many as 79 per cent saw themselves as skilled information workers.

"People who say that Star Office [Sun Microsystems' lighter, cheaper suite] is good enough for them are always pointing at someone else. They don't want to use it themselves," he said.

Staff who spent their day typing one-page letters can get away with using a typewriter, let alone a simple word processor, but more than 50 per cent of word-processed documents exceed 30 pages in length, according to Sinofsky.

"And even if you do only one business plan a year ... that matters a great deal to you. And when I say matters I mean ... we write it like this," he said, writing the final 's' as a dollar sign. "It matter$. You need the best tool for the job."

But, with open source gathering strength, isn't Microsoft simply going to act as a pacemaker for other software writers? Will Linux applications over time not become asymptotic with Microsoft's, matching or even bettering them feature by feature?

"Only if we stay still," said Sinofsky. He sounded confident, which was just as well. The answer certainly matter$ to Microsoft, not to mention the entire US economy it has helped keep afloat for the past 20 years.

Two ways of getting into bed with Microsoft
Microsoft is not about to go away. Even if Linux swept the world, which won't happen soon, it would still be making a fortune on systems, services and tools.

You have only to look at embedded systems - specialist computers hiding in everything from cars to fridges - to realise just what the opposition is up against.

Microsoft has two offerings on this front: Windows CE and Embedded XP. CE has come a long way since its early days, when it earned the all-too-apt nickname Wince.

Like Windows itself it became usable only with version 3.0, and now comes in various guises including Windows Mobile for Smart Phones and Windows CE for smart displays.

The .Net version can pack as small as 350KB and runs on four platforms - x86, ARM, Mips and Hitachi's SHX - allowing Windows developers to apply their existing skills (and Microsoft's tools) to the emerging range of non-x86 connected devices.

Embedded XP requires a minimum 8MB memory, a trivial amount at today's prices. Microsoft has split the operating system into 10,000 components, which are loaded only as needed. But get this: 9,000 of the components are drivers.

So if you are designing, say, a set-top box that needs to run a hard disk, a DVD player and burner, and a Wi-Fi network, you hardly need to write any software: you just pick what bits of XP you need.

Or with XP running one of those routers with a USB printer port: with embedded XP you could, in theory, support any USB device, because you could download any driver as needed.