In thinking more about the business of ranking one OS as better than another I've reached an odd conclusion: specifically, that it's perfectly reasonable to do this with respect to any set of business conditions as long as the evaluation is limited to factors affecting the utility and value of the OSes themselves.
The idea that OpenBSD comes with more, and is more backwards compatible, than Windows Vista illustrates this conclusion quite nicely.
If we limit our discussion to things that pertain to the two OS products sets themselves and envisage usage scenarios involving relatively small machines in personal, research, or server roles then we get a comparison that looks about like this:
| Windows Vista | OpenBSD | |
| Performance | Poor | Out-standing |
| Included Applications or tools | Some games, some utilities, fewer than 50 applications in all | Office Suite, RDBMS, Web Server components, languages -about 4,000 total |
| Security Record | Abysmal | Outstanding |
| Portability | Recent x86 only | x86, SPARC, PPC, MIPS, ARM, others |
| Development risk | Intensely proprietary, winners often become targets | Open Source community licensing and support |
| Network Vision | Replaces local functions | Extends Computer |
| Competitor lockout | Runs Windows software | Runs Unix software, runs most major market Windows software under win4BSD/PC-BSD |
| Ease of Use | One GUI: learning curve starts low but becomes very steep: easy things are easy, hard things very difficult; applications are monolithic, most require re-learning, expertise mostly memorisation based. | Multiple GUIs: learning curve has initial jump, but low slope: easy things are harder than on Windows, hard things easier; applications mostly follow common practices, expertise mostly based on applying knowledge of how things should work. |
| Cost | From $69 per instance before hardware upgrades | Free (optional cost of media for first instance) |
Doesn't seem like a difficult choice, does it?
It isn't - and I don't believe you could construct a comparison based on a reasonably common usage scenario in which all three major Unix variants don't make a Windows choice look absurd.
All of which raises the real issue: why does your new digital what-ever-it-is come with Windows software in a box festooned with pictures of happy Windows users doing magical things? - and, conversely, why is it hard to use the same product with OpenBSD?
My guess is that nothing succeeds like success, or, in this case, that Microsoft has been obtaining compliance at the point of its check book - finding ways to ensure that the big consumer products companies preferentially market its products.
Java offers an answer this - if either the market or the law demanded that consumer products shipping with proprietary software include functionally equivalent or better Java software, then the overall Windows advantage for mass market products would vanish, thereby levelling the playing fields: first for the personal computer and, ultimately, for the office versions.
Makes you wonder how that could be made to happen, doesn't it? A worldwide consumer movement to buy and then return products that require Windows? Civil, anti-trust, or legislative action in the United States? Bureaucratic action in Europe? I don't know, but maybe this is question worth thinking about?