Salesforce-Google collaboration paves way for Web-oriented architecture
Get your Google Apps as part of Salesforce.com, and surrender to your applications in the cloud. The theoretical benefits of the combined service outweigh the negatives.
Salesforce.com's tie-in with Google Apps makes Salesforce the complete center of the user's universe.
But in a new-school twist, neither of these applications completely locks you in. You can get your data out, if you need to (albeit somewhat painfully) from Salesforce, and since you have your Google e-mail stored outside of the Salesforce system, you can effectively leave whenever you want and resplit the applications, should you so desire.
While the technical details are not totally clear, this appears to be an example of Web-oriented architecture, or it at least demonstrates the idea that an abstraction layer allows for data to be more easily integrated. Or maybe it's PaaS (platform as a service)--I am sure it's some acronym.
The theoretical benefits of the combined service outweigh the negatives (mainly clarity around service-level agreements, security, and Google's perpetual beta tests)-at least for now.
This also sticks a big knife in the side of Microsoft, which continues to limp along with its lame-duck approach to the new world of applications. Despite its arguments to the contrary, Microsoft has made minimal progress in its quest for true online applications.
One thing Microsoft would get from the Yahoo acquisition is the enterprise functionality of Zimbra, which could then be integrated into all the other Yahoo properties, along with Microsoft Office.
All of a sudden, the Yahoo acquisition would make a lot more sense, but in light of Microsoft's lack of prescient vision, this seems a long way off. The Salesforce-Google tie-up will likely mean that Microsoft will come back for Yahoo at a higher price, just so they don't totally miss the boat.
I'm not convinced that software-as-a-service applications will take the place of every enterprise application. For example, we're highly unlikely to see high-volume applications jump to the cloud, and we're not that close to multiple-system integration. But as Web-oriented architecture starts maturing, there will be more and more use cases that prove the model.