The Microsoft Dilemma
The issue I speak of relates to how many changes and new features should go into a new release of their software product. If Microsoft adds lots of new functionality, then customers and analysts conclude that the likelihood of bugs within the new code is high and so they avoid product adoption for months or years or until several service packs have been released.
If, on the other hand, the new Microsoft product release is mostly perceived as a refinement of an existing product, with the bulk of changes in the underlying infrastructure or plumbing, then the complaint becomes "why should we pay more money for more of the same?". Whether there are only evolutionary changes, or deeper revolutionary changes to the product, there are complaints.
To be absolutely fair, this problem is not limited solely to Microsoft. Other software vendors face similar concerns once the installed base of their product attains a decent size, or there is sufficient complexity required to deploy the product, or acquisition costs are relatively high.
Microsoft found it necessary to remove features from Windows Vista in order to make the Dec 2006/Jan 2007 deployment timeframe, but it seems that they are moving along with the server release that they can put a few things back into that product. There are a number of critical underlying changes to both the desktop and server editions of the operating system, but this is not the criteria that most people use to determine whether or not to use a new OS. Issues like application compatibility, driver support, performance, and comfort level with the interface are paramount in acceptance of a new OS for most consumers.
Having played with Vista for a little while now, I am positive that there is much to be gained from its usage in terms of productivity. Of course, not everyone will see immediate benefit (and so not everyone need jump on the bandwagon), but many people, particularly those with newer systems and demanding needs, will experience better performance and improved capabilities by moving to the new OS. Yes, it takes a little while for an operating system ecosystem to fully ramp up -- and this is especially true given the huge size of the installed base of Windows XP users -- but it will ramp up, and over the next 6-12 months we will start to see applications, drivers and peripherals that are really designed to take advantage of Windows Vista, both 32-bit and 64-bit versions. There is some really awesome functionality under the hood of Windows Vista that should not be ignored.
As consumers, we have to understand that if we want more frequent OS releases, that they will sport less dramatic changes, and that if we really desire more dramatic changes, that they will of necessity take longer to be developed, tested and released, plus they will be more painful to deploy because of the many differences with the existing product.
Windows is becoming more modular, more secure, and more flexible in many (but not all) ways. This is not a call for everyone to jump on the bandwagon as soon as a new version is released. Rather, it's a reminder that it is often difficult to have one's cake intact while one is eating it. More major changes might make you feel that you're getting your money's worth, but they will represent more complexity for deployment, especially in large organizations. Smaller refinements might seem like less of a benefit for the price, but they provide a smoother path to return on investment.
Continuing in the tradition of Windows Server 2003, I expect Longhorn Server to be a very robust release, with considerable benefits to administration. For instance, Longhorn will finally support the server not having to run Internet Explorer or Media Player by default, which will reduce memory footprint and vulnerability surface area. Sure, there will need to be additional infrastructure changes needed for organizations to reap the full benefit of Longhorn server, but that's not a good reason to avoid deployment altogether. Instead, realize that technology advances are made in incremental fashion, with good planning and timely execution.
I prefer the path Microsoft has chosen for now, which is that major releases will occur every 4 years, with smaller refinement releases in the middle of that cycle. This gives businesses the flexibility to migrate or upgrade when it best suits their organizations, yet still have access to either new technology or tried-and-true technology, as appropriate. And having a more manageable schedule should make it easier for Microsoft to stick to their timelines and not end up biting off more than they can chew.
It should be a win-win situation for all concerned. Let's see how the Longhorn release goes later this year.