This afternoon, Microsoft laid out its next-generation-Windows
Longhorn roadmap, making a delivery commitment and changing how some
new technologies will be delivered. Microsofts new approach to
Longhorn is more staid than all the big bang new version excitement
generated last year. I see the more subdued approach as more realistic
and better suited to businesses that eventually adopt Longhorn and to
developers creating products for it.
The first change is delivery. Microsoft committed to 2006,
something the company may have seemed to do previously, but really
hadn't. Microsoft had committed to releasing Longhorn in 2006, but, in
Microsoft parlance that means “release to manufacturing,” or RTM–also
known as gold code. Real availability could have been 2007.
Today, Microsoft committed to broad availability in 2006, presumably no
later–and, perhaps, much earlier–than for holiday PCs. My August 2003
report, “Longhorn: Implications of the Next Windows Ship Date,” pegged 2006 delivery back when Microsoft guidance indicated 2005.
By the way, Windows Longhorn Server is still tracking for 2007 release.
The second change is about making some Longhorn features available sooner. Microsoft will make some WinFX components–Avalon (the new graphics subsystem) and Indigo
(the new Web services model) available for Windows XP with Service Pack
2 and Windows Server 2003, presumably, with Service Pack 1. According
to Microsoft, developers expressed desire to get these technologies out
sooner. Considering that Visual Studio .Net 2005 is slated to ship next
year and the product is built for developing applications supporting
Avalon and Indigo, earlier release makes sense.
By bringing these technologies to market in, say, second half 2005,
Microsoft and developers could begin releasing Longhorn applications
ahead of the new operating system's launch. That would be very good for
easing the transition and making it less tumultuous than the move to
Windows 95, where consumers and businesses had to upgrade the OS and
many applications nearly simultaneously. The approach is more
realistic, particularly considering how many businesses or consumers
still run older Windows versions. Microsoft could bridge the gap
between the two operating systems, without taking too much focus off of
Windows XP sales.
I wouldn't be surprised if Microsoft released Visual Studio .Net
2005, Windows Server 2003 R2 and Avalon or Indigo components fairly
close together. Microsoft had planned to release a complier as part of
Longhorn, an addition that would increase the integration advantages
between Visual Studio .Net and the operating system. Apparently,
Windows XP Service Pack 2 has a compiler, too. Its inclusion leads me
to believe that today's Avalon and Indigo announcements were plans in
progress months ago.
Release of these technologies earlier also fits in well with what Microsoft calls “managed code”
approach to developing software. Supposedly, managed code is more
secure, and so Microsoft could use these components to further bolster
Windows XP security of applications developed for it.
But, release of these technologies should concern some competitors
or even partners. As evidenced by the Service Pack 2 transition, making
changes to the operating system can cause software applications and
other compatibility problems. Then there are issues of overhead,
particularly with Avalon, and existing PCs. Microsoft had been
developing for 3x today's graphics performance. How many existing PCs
will support Avalon, even if it's released within 12 months.
For competitors, early release of these technologies could be a
grave concern. I had guided some competitors not to worry too much
about either technology, mainly for two reasons: They wouldn't release
until 2006 and, because of the long transition evidenced by Windows XP,
their real market impact would be longer–even four years out. But,
today's announcement changes things for two reasons: Sooner release and
these technologies would come to mainstream Windows, potentially
expediting the transition by a couple years. Certainly, Avalon and
Indigo can offer much to Windows. But, I remain concerned about the
direction and openness of Microsoft's Web services approach and how
that comes to fruition in Indigo. The next six months or so should
reveal how much a threat, if any, Avalon poses to Macromedia's Flash.
The third change ought be good for a few raucous headlines
today, but despite their gloom is no catastrophe. Microsoft plans to
delay delivery of WinFS, which is the so-called, next-generation
Windows file system. Microsoft has indicated a WinFS beta would likely
release around the same time as Longhorn ships. While the press might
make some noise over this delay, I don't see much problem with it.
Previously, WinFS was slated to be ready for Longhorn.
For starters, I don't consider WinFS to be as revolutionary as the initial hype.
Microsoft talked big about WinFS before its developer conference last
year, but, in reality, WinFS turned out to be NTFS with a bunch of new
stuff on top of it. That's not to diminish WinFS' significance,
particularly as Microsoft seeks to improve search through metadata
Additionally, later release is good for developers and Windows
customers. A new file system can wreak havoc on existing applications,
and I imagine all kinds of nightmare scenarios mixing old and new
applications together. Later release should help make the new OS and
new applications transition much easier.
The days of big OS releases are behind us, and I'm glad to see
Microsoft taking a more evolutionary approach with its desktop OS.
Microsoft already does this with Windows Server 2003, for which many
component modules, like Rights Management services, shipped after the
core operating system. A more componentized approach makes sense for
the desktop, too. That's perhaps the best reason for releasing Avalon
and Indigo earlier and WinFS later–and more features to follow them, I
assume. Microsoft would do better by its customers and partners by
shipping components as they're ready, rather than saving them up for
the big bang or holding up a scheduled release to get them in.
Overall, I see today's announcement as well timed and appropriately
toned. Earlier Longhorn hype distracted from Windows XP, which many
businesses and consumers have yet to adopt. Microsoft has committed to
Longhorn's shipment and shown willingness to advance or delay Longhorn
technologies where appropriate. Assuming Avalon and Indigo release next
year, Microsoft would be able to begin early the Longhorn transition
while giving customers more reason to adopt Windows XP. That means
Microsoft and its partners could reap some sales and evangelism
benefits in 2005, without releasing a new Windows version that year. [Microsoft Monitor]