Part of the problem is that anything decent and well marketed quickly attracts basic users. Basic users are fantastic and will talk to friends and encourage others to use your product. You can’t really win anything large without the basics (niches are about the best you can hope for otherwise). The snag is that the basics can’t help you with the engineering and they are numerous and very noisy. Thus if you attract basics in numbers you can’t handle your open source product will see a wave of support requests but have no one to actually work on them. Eventually an important basic will come along, see the problems and damn your development for good (people can only put up with so much for so long).

Another problem is that there is a very vocal set of people who just want Windows but free. They do not care for any difference and will complain upon finding one punishing developers for not having the Microsoft attitude and caring about users. Some of these people should be buying Windows but choose to rip it off and are not willing to put up with an inferior alternative.

One massive problem are patents. These are simply going to stop a large part of the open source experience being smooth and easy. All you have to do is look at Mac users who are saying “Well your experience pales compared to ours. We can do everything you can do but with less effort”. Apple has serious IP, can (and will) license those patents for what it needs to do and has spent a larage amount of time making a usable core which means their users DON’T have to take time. The idea of spending time forcing OSX on to non Mac hardware is unusual as you are just trading time for money.

Finally people hate change but backwards compatibility is expensive and boring. Unfortunately there is a very high cost to adding new features to existing products without breaking things. If important things on your open source desktop are actually closed you will find your hands eventually tied. This is because open source systems make radical changes that support new functionality and will change most of their existing stuff to support the new stuff. The bad news is that this doesn’t gel with binary drivers (that expect unchanging interfaces) or a system that continually receives new additions (and security updates) throughout its 8 year life. Each clean break sacrifices more users…

- Someone who cares about the BSD subsystem in OSX.

At a recent show I talked with the FreeBSD booth several times. Each time I got more interesting information. Keep in mind that several years ago I predicted the extinction of the Chevrolet Suburban and FreeBSD in the same breath. FreeBSD is still around and growing.

ZFS is native in FreeBSD 7.0. FreeBSD is Unix. I am a Unix bigot.
I wanted Mac OSX, in particular the Aqua layer, for Unix. Didn’t happen until someone convinced Steve Jobs that was the way to make his dream, Lisa, live. Or maybe it was the other way around?
The rumor has been that OSX is built on FreeBSD. Apple says only selected parts. The heart of OSX is the Mach kernel according to Apple’s official statements that I have read. Who knows? Who cares?
Apparently FreeBSD enthusiasts.
If Steve Jobs permitted even a minimal Open Source port of Aqua then FreeBSD could sweep both the Open Systems server market and the desktop.
The recent changes required to move from the PowerPC to Intel moved Apple very close to having this.
The desktop would provide FreeBSD fans the revenue stream to finance the server market expansion. Free only goes so far in the server market. Takes “Big Bucks” to gain market share.
There was kind of a twinkle in the eyes of the FreeBSD guys. The twinkle that says you may be more correct than you imagine but the possible legal retribution keeps us silent.
Just “Food for Thought”…

It isn’t easy to get product requirements right by listening to customers. Data center folks are typically operating at about the 5 foot level. When they have a problem it tends to be a very “here and now” specific problem.

The hard part is deciding if the customer’s problem IS the problem, or the symptom of a deeper problem. For example, volume management is a problem. Now we can make the volume management interface better, automate common procedures, design better menus and the like. But the real problem is that we are managing volumes instead of pools of blocks.

Marketing vision only goes so far. Marketers don’t know what the smartest engineers know. What marketing *can* do is help figure out how attractive an engineering innovation is to customers. They can also dress it up and make it as attractive as possible.

The worst situations are when product management starts trying to engineer the product and engineering starts trying to market it. Chaos for sure.

The bottom line for desktop Linux is that with some modest exceptions all the marketing is done by engineers. As a result the product isn’t competitive with Windows and OS X for the huge majority of users.

Robin

I agree with a lot you said. It seems to me that geeks simply don’t get this desktop usability business. How make desktop applications more consistent, i mean similiar menu, same shortcut keys for cut-n-paste and such.I still can’t get used to ctrl-shift-C in terminal and ctrl-C in firefox on my Ubuntu…. I read about the road map for next Ubuntu, it’s no where on the radar screen. Linux distro has to learn a lot about Mac if they’re serious about pushing penguim to the desktops.