MS Analysts The theme for Microsoft's financial analysts meeting last week was "the road to the .NET" but it still seems that real plans are sketchy, and the paradigm shift that it is supposed to represent is still being invented. Gates' speech was not exactly the stuff to excite financial analysts; far from pitching them exciting new ideas as to how Windows might evolve, the chief software architect started by suggesting that his 1976 vision had come true.
Software would be important. The platform would be significant. Hardware would advance. The stuff of visions, indeed. Gates then moved on to a reiteration of favourites from earlier speeches: that most information is still stored on paper; that broadband is important; and that the digital dashboard was "the first step down the road to .NET" - something of a retrofitted claim.
For Gates, ".NET is very analogous to the founding vision of the company - the PC is the empowering tool; first Basic, then Dos, and then Windows". He went on to suggest that "the Internet really becomes a platform" when messages talk to each other rather than being presented on a screen. Fortunately he amplified this observation: "Now, you could say, well, isn't that just client-server? ... And yes, we did, but it was in a very constrained environment." The idea of computers working together across the Internet, he said, required a different way of doing the programming, which he called "loosely coupled".
Gates made it clear yet again that his personal interest is firmly on the consumer side - digital photos, TV, movies, music - rather than on business. He updated this limited vision with the notion of a screen on the refrigerator, connected by wireless and running "a specialised version of Windows" - but unfortunately he kept the specific purpose of this innovation from us.
Keep talking the tablets
He claimed that the tablet device he has been pushing for years will "replace the portable market" and achieve a "substantially higher volume". Of course all portables will continue to advance - and probably in ways that do not always suit Microsoft, so far as the operating system is concerned - but his apparent view that his dearly beloved tablet will be a notebook killer is absurd. We were also amused that Gates suggested that there would be meetings at which people would be pointing their tablet devices at each other and beaming contact information: he may not know that this infrared handshaking has been a reality with fanatical Palm owners for some time.
Gates did have a sage observation on the nature of the consultancy business: "The interoperability problem has been around for a long time. How do you get these systems to work together? In fact, you look at the consulting business, probably half of the time they spend is just glue code that makes things work together."
<$200 a year for services?
So perhaps the real point of Microsoft's .NET initiative, besides changing Microsoft's financial model in the hope of increasing revenue, is a reaction to the fact that software, Microsoft software in particular, has become too complex for the average user to maintain and use to build applications, so that has to be outsourced. Gates volunteered a figure for how much he thought should be charged a year for each "knowledge worker" to get software from a service: "it's worth even more than, say $200 a year to get the latest service..."
Ballmer: more cash out of consumers
Steve Ballmer reinforced the money side, commenting on the future of personal subscription services: "Today the amount of money we make selling software into the consumer market is much, much, much, much, much, much smaller than the amount we make selling software onto business PCs. And all I see by being able to offer people services that they subscribe to, where we take care of their machine, we give them the software, we update, we take care of the hassle and complexity that you have to do sometimes when you install software yourself, not that we're not going to run smart software, but we'll take care of the installation, we'll manage it..."
So far as how Microsoft envisaged .NET services would be run, Ballmer said in answer to a question that he foresaw that the bulk of services "would be run by the guy who wrote the code", with third-party ASPs being involved as well. Microsoft itself might run an Exchange service or bCentral, probably with a third-party hoster like Exodus, although "we take the full responsibility for the customer".
Muglia: stop them stealing our software
Bob Muglia, vp of the business productivity group, mentioned that another reason for the services model was to cut down on piracy by having electronic control of users. He advanced the view that there was a higher rate of piracy in small businesses than in large ones, but the theory was based on lower sales to small businesses. Perhaps it had not occurred to him that canny small businesses had not swallowed the propaganda that they needed constant updates. It is also being whispered that the primary reason for a downturn in software sales in some sectors might not be Y2K, but rather the result of a greater general realisation that upgrades are not particularly essential.
Ballmer didn't exactly accuse small businesses of piracy, but he did comment that "The small-business market is not a very active market. We get something less than 25 per cent as much revenue per computer in the installed base for small businesses versus large businesses. Small businesses are loath to upgrade. Small businesses pay people a lot of money to get their desktop applications integrated with their accounting and their customer management, and then they don't want to change anything."
The long hard code to .NET
The next generation, Gates suggested, would have the Internet as the platform, with the software being .NET and consisting of XML, being scalable, and being a service. It was a bit of a thin vision, but that's all he had on the slide apart from the "Empower people through great software any time, any place, and on any device" mantra. The next slide confused the issue because it was headed ".NET: A Platform" and gave the new terminology that Microsoft would be using in transforming applications to services. The user interface becomes the "user experience"; a compound document is a "universal canvas"; a file system would be an "XML store"; message queues would be "XML messages"; APIs would be "WS, Building Blocks"; and the Windows Platform would be the ".NET Platform".
The .NET development would stretch over "many, many years" Gates said, and "the road to .NET has many big milestones on it. The next release of Microsoft Windows will have some significant aspects in terms of storage and authentication, and instant messaging capabilities. The next version of Office will have very significant things. The enterprise servers is where we've made the most rapid progress by embedding very rich XML support, even in the servers being released this year. The generation after that is when you get the 100 per cent capability." His roadmap suggested that in 2000/2001 there would be Windows.NET (Whistler), Office with .NET, .NET enterprise servers (i.e. BackOffice and DNA 2000), with MSDN.NET also being released. For 2001+, Gates listed Windows.NET (Blackcomb) and VisualStudio.NET.
There would be one code base for Whistler, which would include Datacenter Server, Advanced Server, Server, Professional Client, Personal Client, and Embedded releases. It is hard to see how optimisation would be possible for a single code base intended to scale from reading a smart card, through playing a game to running an enterprise. The truth is of course that it cannot, but it will be interesting to see how Microsoft intends to explain this away. It is still even a completely open question as to whether this direction will prove to be the right one, although with so little being defined, it would be easy for Microsoft to map the road as it staggers along it. ®