It doesn't matter how good BeOS may have been, its fate was sealed by the decision to create a closed source operating system. On the Fortune Global 500, IBM is #19. Microsoft is #175. Yet IBM's OS/2 failed to compete with Windows. Why? Partly it was their own fault, but Microsoft's OEM-practices (which were even worse back then than they are now: PC makers had to pay the "MS tax" even for systems without an OS) made it extremely hard for any competitor, no matter how large, to gain ground. The nature of operating systems complicates things further: Since everything else depends on the OS, switching the OS is a lot like trying to "switch" the foundation of a house. You have to tear the whole damn thing down and rebuild it to do it. No wonder that most people didn't see a reason to do it on their own.
Be-founder Gassee probably finally understood how the MS monopoly works when Hitachi shipped a notebook with Windows and a disabled BeOS partition. They had agreed with Be to ship the two-OS system, but buckled under OEM contract pressure from MS and made the second OS unusable shortly before shipping.
The only OS that can compete with MS is a completely open one. For the first time, MS is facing a competitor they can't beat on price. A competitor without a name: There's no company that can go bankrupt for the open source idea to die. There's nobody they can beat up. It is essential to understand this difference. If you commit yourself to a closed system, you can only lose. You lose if the system dies, and you lose if it wins because then, once again, you're locked in. The only situation where you don't lose is when the system remains active on the fringe, such as Apple, which has made a deal with MS: You port your apps to our OS, and we don't try to steal your market share.
But with an open system, you cannot lose. You win if the system grows, because then the closed system gets replaced with an open one. But even if the system shrinks, you still don't lose, because there's always a way to migrate -- you could even take all those open apps and use them under a closed system if the need should ever arise.
For the time being, Linux is in the spotlight, but something else might be built from open source components. It doesn't really matter what it's called, it only matters that it's open. Be doesn't deserve your tears -- had they won the impossible fight against MS, they would now be in MS' position, and be subject to the same corporate politics.
Building the open desktop OS
Linux still has some catching up to do in the client area. It was only a year ago when in literally every Slashdot discussion about Linux on the desktop there was a consensus that it "wasn't ready", and many thought it would never be. Just this month Red Hat has released their first desktop distro, and United Linux is planning to do the same -- because now they feel that major components have become available that are necessary for corporations to use Linux clients in some areas. In other areas like games, it is still lacking. But it's getting there.
I'm amazed, for example, by how far WINE has come. Crossover Office, based on WINE, allows you to run Office, esp. Word, almost perfectly. Even IE works like a charm. If this currently proprietary stuff finds its way into the open source tree, I'm sure WINE will soon be able to run most smaller apps without problems. This will be another major puzzle piece for Linux' client competition.
As for the database file system, yes, it's obviously a killer feature. Even MS has realized this and plans to do something similar, or the same, in Longhorn, but this is still foggy and far away. I'm not sure if Linux will get it before MS will. I surely hope so, but this is a large effort that needs to be coordinated in the same way as big projects like GNOME and KDE. There needs to be a DBFS project that covers everything from the kernel modules over command line interfaces to the high level UI stuff in Konqueror, Nautilus etc. If this isn't done in a coordinated fashion, it will end up as crappy as the clipboard currently is under Linux/X. [Note to zealots, do not correct me on this, or I will get angry.]
That's where closed efforts like Be have a huge advantage. They started from scratch and could come up with a good, clean design. Linux as a whole grows in an evolutionary fashion, only subprojects are engineered.
But I'm sure that if the whole DBFS thing gets widely acknowledged as the killer feature it truly is, efforts will quickly start to copy it in a 1:1 fashion. FWIW, Ximian Evolution has full text indexing of all mail; I can search 10000 mails in a couple of seconds and V1.2 will have an even faster engine. I could never do this with any Windows mail client I tried. For email this is already pretty good, and for text files, grep is usually fast enough, even recursively. There are some indexing tools for local content searches, but I agree that stuff like that is best done on the FS level.
Copyright law is bad: infoAnarchy · Pleasure is good: Origins of Violence
spread the word!