There was a time when I was excited about new operating system releases. Lately, as in the last several years, I have dreaded each new release.
With Linux, I know that features are going to be taken away, more bugs will be introduced, and I generally know that configuration files will be increasingly harder to read. I also know that the entire software stack will become more and more over-engineered. Meanwhile, long standing problems will not be addressed, and even the oldest of communities are going to have crap tons of infighting that will impede development and release. Worse, I know that many older programs will not run on newer versions of the same distribution, and I may have to pay some third party to port the applications I need.
With Windows... Vista? Windows 8? Really? Windows XP was hated when it came out. People only began to like Windows XP because it was on the market for a very long time. They just got used to it. Realistically, Microsoft should have simply rebranded 2000, updated some of the unseen bits, and sold the crap out of it. Windows 7 was only successful because it wasn't Vista. 8 and 8.1 were hated by just about everyone. I don't really expect 10 to be good, I only expect it to be better than 8 and 8.1 when compared to them alone. Windows has been so bad for so long that Microsoft as a brand has suffered.
With OS X, I know that the OS itself will either have new features or it will be more stable and more performant than the previous release, but iOS 8 has been worrisome. The bugs and performance issues (while admittedly not as bad as Android) have been conerning. It's very unlike Apple to release half baked shit. I know that while new features will be introduced, the features upon which I rely will likely not be taken away. Unfortunately, I also know that things will be increasingly dumbed down outside of the OS (Final Cut Pro, Aperture, and likely more to come). I know that hardware be increasingly more integrated and non-upgradeable. I know that if I want to see something more powerful than an iMac and cheaper than a MacPro, I will have to build a hackintosh.
There is no longer an OS that is objectively a good choice. For the most part, people can only choose what sucks least for them. Do you need to do professional creative work? We used to say "buy a Mac," but now it's more like: "which OS do you hate least?" This is terrible. When people are only using your shit because they hate everything else more, you have a problem. Of course, if you are an insanely talented developer and business person, you have a market opportunity.
Mainly, I feel that we all need to slow-the-fuck-down. You do not need to make releases every six months. You cannot make releases every six months. First, people do not want to do an install every six months. Second, unless you have some super human staff, you cannot possibly have developed and tested something as complex as a modern OS in 6 months. That same staff is not going to be able to support 15 different versions of your OS in a few years either, which means you are going to alienate much of your customer base over time. Not to mention that OS development is expensive to both the developer and the buyer. OSes shouldn't need to change until there is significant hardware advancement that was unforeseen when the OSes were written. 64 bit? That validates a new OS release. Thunderbolt? Depending upon how your OS was written, you should be able to support it with a driver release.
As for major software changes, is the feature you want to implement going to alter the OS in such a way that every part of the OS is affected by it? If yes, make a release. If no, just issue a fucking update. This isn't rocket science. You should be focused on securing and stabilizing your current offering, and the development of a new major OS version should only be on the road map when significant changes occur in the computing world at large.