Tomorrow Bill Gates (convicted monopolist) and Ray Ozzie (respect) are expected to announce the third coming of Microsoft. Just as they turned a massive organization on a dime to embrace the Internet, they will break with the past to offer software as a service. Initially it may be a storage offering, communications package for SMBs and CRM because of competitive pressure, but this pressure is from all sides.
The difference this time is it requires breaking something, not venturing out into whitespace. Steve Gillmor may be right that they are discounting the defensibility of the Office empire. But if they shift the value to serviced software, they also directly undercut their VARs, the distribution muscle that profits from overcoming the complexity of bad software.
As an entrepreneur with software as a service as an important part of my business model, I'm not worried in the slightest. Initially it is FUDing out the next rev of the web. But then the task is making stuff simple because the burden shifts back to them to service it. And with this move the game is on, on the turf of other credible Giants. Sergey Brin wasn't wrong when he said at Web 2.0 that Google Office, with the disclaimer that it may not exist, wouldn't be anything like the old Office.
You see, replicating Word for the web isn't social computing. Office and PC era apps do a great job serving the lone producer's need to complete finished documents. But the process of creating that document is only served through clumsy means (read: email). The next generation of productivity gains will come from something new.
I'd also make an argument that the business model for software is shifting to a blend of software as service delivered as hosted or appliance with the production dynamics of open source. At the end of John Markoff's book on the birth of the PC era, he juxtaposed the freedom to share sought by Hobbyists vs. the profit motive of Microsoft as the defining conflict. As that era comes to a close, a compromise has been found in commercial open source.
I think Microsoft's turn was first made when they readdressed the flaws in their security model. They had embraced the Internet, but with a weakness that today provides a negative user experience. Spam, spyware, adware and viruses plague the PC and Microsoft is held accountable for it. The opportunity they realized is that trusted computing models can put them back in control of the network. They will have their own DRM'ed lightnet in which any competing software as service or unsanctioned content vendor will have to get through them to the desktop. Not just through the web, but as deep as P2P itself.
Jerry Michalski said something brilliantly simple the other week at a collaboration conference. That we are certainly not waiting for Vista, we have lost trust in these guys and it's time to get these things out of the OS.