The five-year long march to developing Windows Vista exposed just how broken Microsoft’s Windows-development process is. It took far too long, major features were dropped along the way, and it was wrapped up in more red tape than the IRS. That’s why I think that Vista will be the last of the “big-bang” Windows releases.
Mary Jo Foley has an excellent analysis of what went wrong during the Vista development cycle. As one example, she cites the fact that it took more than 40 programmers a year to deliver the the final Vista Off button.
Overall, the problems are obvious: bloated staffing, too many levels of management, and far too much bureaucracy. Microsoft has become the slow-moving, bloated company it vowed it would never be.
Everyone I’ve talked with at Microsoft recognizes the problem, and I think they’ll take care of it. The next version of Windows, the company claims, will take closer to three years than five, and will be a far more incremental upgrade than Vista was. In fact, it sounds as if Microsoft has all but given up on its attempt to remake the underlying file system into a true database.
The next version of Windows will feature incremental changes rather than dramatic ones. But I’m not entirely sure that after that, there will even be Windows upgrades in the sense that we now know them. Rather than upgrade the entire operating system, I expect Microsoft to instead offer online-upgradable modules and feature sets. In this model, you wouldn’t ever buy an entirely new version of Windows (except on a new PC, of course). Instead, you’d buy a subscription to Windows, and pick and choose which new features to install.
This would not only allow Microsoft to simplify development, but it would give the company a steadier revenue stream, via subscriptions rather than operating system upgrades. The change is inevitable. We won’t see any more massive Windows upgrades as we’ve seen with Vista.
Two comments:
1. Joel Spolsky wrote those comments about the obvious problems with Microsoft.
2. I think that your comments that the next version of Windows will be a subscription and be more web-based will be true.
How could that work though? A pick-and-choose system would create a software equivalent of the vast variety of PC hardware, which while advantageous in some ways has been a huge headache for MS and technical staff to support, and lead MS to try very hard to limit and control the PC hardware architecture. An explosion in the number of possible software component configurations in the base OS is a support, stability and security nightmare in the making.
Unfortunately Linux already has something like this already, in the form of the tens of thousands of packages available and their dependencies. Package X requires packages Y and Z plus verison 3 of package Q, meanwhile package K requires version 2 of package Q and doesn't work with version 3. There are package managers that try to simplify managing this, but it's a big problem.
Suppose MS brings out a fancy new filesystem. Is it a pay-for package? What applications will depend on it? Will they work with plain old NTFS, and what happens if MS decide to base future OS features on an optional package?
There's a name for optional components you ad to an operating system, and it's 'applications'. Microsoft's fragmentation of their operating system into half a dozen flavours may look good in a marketing presentation, but in practice it looks like a huge blunder on technical grounds. Unfortunately, the marketeers are in charge over there now and further market experimentation with new sales and support models isn't going to improve things.
I think what they'll do is lower the initial price for Windows but then charge people every year to renew the license, keep getting security updates, etc.
They'll claim it's what their customers want, and that it makes Windows more affordable to people, and who knows, maybe they'll be telling the truth :)
But what it will really be about is continuing to make money off of Windows buyers for years into the future, rather than the big "pop" at the buy and then *nothing* - which forces them to have to always develop the Next Big Thing to keep the money rolling in.
This way, with subscriptions, the money will just KEEP rolling in, year after year, as long as people don't want their Windows to deactivate - they'll kind of have people by the proverbial testes.
The counterpoint to this, of course, is that Microsoft does continue to provide security updates and add-ons some years into the future, so maybe they should be compensated for continuing to do that. Otherwise, why shouldn't they discontinue security updates to Windows 2000 tomorrow?
But really, what it will be is about monetizing all those many, many, many people who get one version of Windows and then never upgrade because 1) they fear change (earnestly), 2) they just don't see the value in upgrading (i.e., what they have is "good enough").
Hmm... Yearly payments to Microsoft to keep your Windows functioning - maybe we can call it the "Microsoft tax".
Sounds simple put this way! It ignores the complex issues of systems that do not interconnect in networks: what if you are required to run in a secure unconnected environment! What then? You seem to blissfully ignore this very real world situation!
vista is a security downgrade, they messed all things up, the instalation of vista+drivers takes ages, wtf .. wasting resources like ram, disk and others for no reason