It’s almost exactly three years since I first saw (under strict non-disclosure) a demonstration of the new Microsoft client operating system that would one day become Vista. I was enthusiastic back then, and I’m just as enthusiastic now – and not just because I’m tired of the fragility of an aging Windows XP. Throughout the history of computing (not just the PC, but all computing), hardware performance has improved much faster than the software that makes it possible to use. As a result, we have seldom been able to leverage the full power of the computing platforms we use. That’s especially true today, where new client-side architectures like PCI-Express, 1000BaseT Ethernet, USB 2.0, FW800, SATA and powerful graphics subsystems could potentially free us from the 1990s limits of the original PCI bus, IDE interface and mostly slow-speed serial communications.
We won’t get to this potential without a new OS. Windows contains just too much legacy support to be efficient and stable when stressed. There are too many APIs; too many ways to do just about anything; too many ways to write poor code (especially poor driver code) and get away with it. It’s past time to start again, finally sacrificing unconstrained backwards compatibility for greatly enhanced future capability. Giving up some increasingly irrelevant compatibility guarantees that we can get the stable, secure, adaptable software platform we will need for the next billion PC users.
Sure, you can buy OSX now; it’s a great OS (I use it regularly), but it’s relatively expensive and locked (at least for now) to a closed hardware platform. You can try Linux (I use that too from time to time), but frankly I don’t like any of the available UIs much and I don’t see it keeping up with the evolution of hardware either. And sure, MSFT should have fixed all this with the XP wave, but the risks back then (XP started life in 1999, remember) must have looked horrendous to MSFT and they (justifiably) opted for evolution rather than revolution.
Not this time. But restarting from scratch carries risks of its own. Re-architecting a complete software platform based on what we (thought we) knew in 2002 would likely be available after 2004 meant placing some difficult bets, while having to support the entire global technology ecosystem that depends on the WinTel duopoly isn’t easy either. No matter how good the team is, there were bound to be some false starts and blind alleys. The important thing was (and is) to keep the dead ends out of the final platform, and that takes time.
So, as much as I would like to have Vista today, I would counsel patience and suggest that instead of climbing all over Microsoft for being late, we applaud them for taking the time to get it right. By the criteria I use to gauge an OS, Vista will be worth the wait, and all those hardware upgrades I have been accumulating for the past few years will finally start to deliver on their potential. I can hardly wait, but I’m going to.