Legacy
I am selfish. I try not to be. Selfish Phat wants Windows to work.
In that part of my life where I pay the mortgage (yes, I still can), I do the thinking for the enterprises that cannot. I'm a "special needs" consultant.
The majority of my clients bring a Windows Server legacy with them. Regardless of what anyone says, most are either incapable or unwilling to mix platforms between their servers and their workstations. And no, Linux cheerleaders, they won't abandon their Windows server installations just because you think Linux is cool. Remember, the CEOs get together and play golf. They talk about work. Linux has not made many CEO friends in mid-to-small-business. Sorry.
The desktop legacy also follows Windows. The custom interfaces, server tools, enterprise applications, and enterprise software I build and support is in a Windows environment -- from server to data store to desktop to whatever.
The challenge in supporting or migrating operating systems in small to medium sized businesses is not limited to the operating systems. It extends to the enterprise-level software, both vendor-supported and customized, that runs each business. Even the small(ish) company with fewer than 50 employees has, over the course of ten years, invested millions in the purchase, customization, integration, maintenance and modification of that enterprise software.
A migration from one "brand" to another would be prohibitively expensive.
As a reminder, the behemoths are still chained, unwillingly so, to millions and millions of lines of COBOL code. They cannot move for the very same reason the small and medium sized guys cannot move from Windows Server and workstation software -- the very core of the business exists in an uneasy, dependent, symbiotic relationship with legacy code that is usually more than seven or more years old.
These companies, the backbone on which the economy moves, must be stable. They must move slowly. They must invest wisely and incrementally. "Free" operating systems would demand a re-investment in millions for a single small corporation to re-implement their ERP, integration, communications, and all the other forgotten automation that just runs.
This is why a 40-employee company cannot slap-dash drop Windows Servers and Desktops in favor of your particular flavor of open source something. The business does not run on Linux or Ubuntu or Red Hat or even Open Office. It runs on the core business enterprise applications that have been bought and built and customized and modified over years of trial and error (and more programmers and IT geeks than anyone can remember). And nobody kept good notes.
I would argue (in court if necessary) that Microsoft, then, has a fiduciary obligation to provide these companies, and all companies, with competent, usable, stable, efficient platforms on which these companies can run their business/enterprise applications.
Vista does not, and cannot fulfill that obligation.
I'm not banking on WineDOS 7, either. If history is any school, Microsoft has already screwed the pooch on that one. In similar fashion, every Microsoft release of applications or operating systems since 2003, save one, has been a bloated, resource-crushing disaster. (The jury's still out on SQL Server 2008, but I'm wondering why I need it when SQL 2005 is still on the shelf?)
Regarding the servers, Server 2008 is decent. It made some progress over Server 2003. I have installed Sever 2008 on some new servers, but I haven't replaced any Server 2003 implementations with it.
And the workstations, which is the point of this monologue, are all still XP Pro. Why? It works. Vista doesn't. And even that mythical under-fifty-employee company I reference would have to spend tens of thousands to install (notice I could not bring myself to write "upgrade?") Vista even on the 40 or 50 machines they use. (They would probably be required to buy 40 or 50 new machines just to do it.)
It is all mathematical. Microsoft ***MUST*** produce and deliver a workstation operating system that is affordable, usable, dependable, and works with their current business/enterprise applications.
Whoops! They already have. XP.
Companies must have clear, convincing, low-risk, low-cost, usable, workable, easily-integrable options in everything. This is especially true of infrastructure. There will be no mass-migrations from Microsoft at any company. Any new machine or operating system or desktop applications must be inserted into the business management process like it was tooled and oiled and installed by a master. Otherwise it will be disruptive, expensive, risky, and bad for business.
This is why Vista will not succeed. This is, too, why anything besides a Windows solution at a Windows-based company will not succeed.
If Microsoft fails to deliver a success with Windows 7, there will be more businesses in distress than some little high-risk mortgage crisis can cause. If Microsoft fails to deliver a success with Windows 7, tens of thousands of companies will be forced to start all over -- and many of them don't have the margins.
You and I, geeks and experts, can install Linux or What-U-Too on our desktops. Some can afford to upgrade to Macs and OSX. I have started installing Server 2008 on my own desktops to use as my workstation operating system.
Businesses are not so flexible. They do not have the overall depth of expertise that you and I retain. They are at the shallow end of the technical pool, and that is how it will always be. (It will get even more and more shallow, relatively speaking, as technology progresses, too.)
Extending the support lifeline of XP is the best news for business I've seen in a month. But I think Microsoft is not doing it for altruistic reasons.
I don't think they're ready to fend off the mobs should XP die and all Microsoft can offer is a Vista monster, or its Windows 7 offspring.
I think Microsoft is buying time.