MS Win2k-based consumer OS rides again as Millennium fades
Windows roadmap loops back a year after Win9x-based excursion
Microsoft now seems to be confirming the strategic change we read into the company's latest reorganisation announcement on Friday. By merging the two Windows divisions, Consumer and Business, into a single unit within Jim Allchin's Platforms Group the company was effectively signalling the end of the Millennium consumer Windows project, and reverting to a 'one Windows' strategy, at least for the PC space. Millennium has been in development for most of this year, and made Beta 2 late last month. But it has also been the subject of a tussle within Microsoft; earlier ambitious plans for the OS were steadfastly pooh-poohed here months ago, and subsequently the project declined to the status of a service pack (not entirely on our say-so, we concede). Microsoft is now being quoted in today's Wall Street Journal as saying Millennium will go out next year as "a relatively minor upgrade to Windows 98," and is also worth noting that MS is telling the WSJ that it will be previewing "features" (no, we're not sure what that means either) of Neptune, the consumer version of Win2k, shortly. While we're sifting the spin coming out of Redmond we'll just draw your attention to something you may have missed - an interview with Jim Allchin in the December issue of Wired. This would have been conducted a few months back now, when any internal battle over Millennium would still have been decidedly live, and so its claim that "Even the Dos-based next-generation consumer OS - codenamed Millennium - has been pushed back to 2001" is intriguing. The piece doesn't attribute this claim to Allchin, and it could be a typo, but one wonders... Anyway, you'll have noted that the Microsoft strategy for consumer Windows now seems to have rewinded a year. Until early this year the Microsoft plan, as stated by Bill Gates himself at WinHEC 98, was to cease development of the Win9x codebase and follow up the shipping of Windows 2000 with a consumer OS based on Win2k. This plan was shelved shortly before WinHEC 99, and the dual-track Millennium-Neptune programme were kicked off instead. Neptune was intended as the new OS on the common codebase, but was clearly headed for a ship date in 2001 or later. Millennium's role was therefore to fill the gap. Note that the Wintel strategy papers that were live at the time of WinHEC 99 still implied the appearance of a Win2k-based consumer OS for the next generation PCs due to ship in autumn 99 (i.e. now), so a hasty revival of the 9x development programme seemed necessary. The trouble was, of course, that doing much useful in terms of software for easier to use and legacy free platforms would be hard, and even if you weren't doing it (as turned out to be the case with Millennium) the OS wasn't going to make it out of the door for Q4 99. Which is why more recent versions of the PC2001 design guide have appointed Windows 98 SE as the OS for these platforms instead. That leaves Millennium looking a bit pointless. As a minor rev of Windows 98 it could just be a service pack and/or an OEM code refresh for around mid-2000, but it won't butter many parsnips either way. Microsoft can use it to introduce some of the new features it promised in the EasyPC initiative, but it could equally have done so without the more radical up-front hype. Win98 SE was no big deal, but added some stuff, right? That's an indicator of what Microsoft is likely to do with Win9x for the future - minor revs to the OS will act as a kind of holding action until the company is able to ship the main event, Neptune. And there's a possibility that the promised Neptune feature demo may signal an acceleration of development - at the very least Jim Allchin probably thinks he can accelerate it. Allchin, you'll probably recall, has been driving Windows development for quite a while and has had a lot of trouble with synching twin-track development (IE for Win98 and NT, for example). As tear-stained trial emails show, he's positively allergic to this stuff. One of the earlier suggestions for Millennium was that it would introduce next generation UI technology; this would have been something that made the OS look more radical, and it would also have been another one of those headaches for Allchin, who'd have had two different development groups trying and failing to keep in step again. But the new consumer UI developments now seem to have departed to Rick Belluzzo's Consumer Group along with former Consumer Windows Division head David Cole. So Allchin has a cleaner sheet than he's had for some time. That doesn't necessarily mean that Neptune development is going to be any less of a pig, but it does mean the Platforms Group can assign a lot of resources to it. And if we're talking a ship date of say, mid-2001 circumstances should have changed by then. Consumer PC hardware specifications will have gone up, and the Win2k code will have got tidier and more efficient. A lot of legacy games support will have to be lost along with legacy Dos and hardware support, of course, but in 18 months time, Microsoft possibly calculates that it'll be able to sustain that. And if not, there will always be the 2001 Win9x service pack ready to plug the gap again... ®
Sponsored: Fast data protection ROI?