Windows Stifles Innovation

Windows Is So Slow, but Why? - New York Times Windows is indeed stifling innovation — at Microsoft. When I was at IBM, I worked with various people involved with MVS development. Most of the time I thought it was pretty cool that there was this operating system that was (then) nearly 30 years old and that customers frequently were able to run applications developed in the 1970s on the latest releases of MVS. But a 30 year old software environment gets a lot of cruft. Bugs get so intractably embedded in the routine processing that you can't just remove them as various other parts of the O/S as well as applications atop the O/S become reliant on the existence of the bug or misfeature.

Microsoft has staked out a position to be as backward compatible as possible when upgrading Windows. This means that every hack, feature, undocumented internal hack and feature has to be supported in new releases, making the development of a new version exponentially more difficult. Where Apple has had no qualms about telling users of applications reliant on older features tough noogies, Microsoft is addicted to its user base. With good reason since if you make it too difficult to upgrade, users will either not upgrade or consider alternatives. As long as you're going through the pain of change, why not go further and drop the reliance on Microsoft?.

In the 1970s, IBM set out to build Future System. Depending on whom you talked to it was going to be RISC based, have all of the latest ideas and improvements (circa 1970) in computing. It never launched, though various ideas from the project were reused in the AS/400 and other products. Again, depending on whom you talked to, it set IBM's mainframe O/S development back years.

This is not the first time Microsoft has poured energy and resources into developing a new release of Windows. In the 1990s Microsoft embarked on an ambitious plan to succeed Windows NT with the Cairo release. Reading a feature list for Cairo is eerily like reading the feature list for Vista (an object oriented filesystem was promised for Cairo, WinFS was to be the Vista version but was dropped in the last re-jiggering of the Vista release plan).

The problem is you get a certain inertia, the bigger the change to the environment, the more features you want to add to make it palatable to customers. You get into a vicious circle. What surprising isn't that Microsoft has hit this point, but that this is the second (or third) time in just under a decade that it has committed, then gradually backed off from, a major operating system upgrade.

Posted in