The End of Open-source?

Six years ago I speculated on one factor that could spell the demise of open source. Here are some more.

I've spent the last year looking for the perfect Linux desktop; and concluded not only that it doesn't exist, but that it never will, nor ever could. In the last 10 years, I've used Fedora, Ubuntu, Crunchbang, Xubuntu; I've tried Debian, openSUSE, CentOS, Scientific Linux, and Mageia. All have drawbacks, none are immune to crashes, freezes, or other irritating flaws. So while Linux works very well as a server, as platform for a desktop environment, it simply can't compete with OSX and Windows. In part that's because I still need to use products like Excel and Word which almost run in Wine and do, sort of, in Crossover. But it's also a function of the fundamental design flaw in the production process by which open-source is created. 

Just as Economics as a field has benefited hugely from having a broadly unified paradigm, so have Windows and OSX. If something needs fixing it gets fixed, generally relatively quickly; in the open-source community it's done when someone has time. And since contributions to open-source are motivated more directly by self interest (itch-scratching and status within the community), the connection between the market and the production process is weaker and less strategically oriented than it is in the proprietary world. What gets fixed in open-source depends not on what the as yet unserved users might need but on what existing users want. This is akin to Clay Christiansen's "Innovators dilemma"- serving current users rather than the larger potential market.

Without a unified guiding vision and a single roadmap, efforts are unfocused and energy is wasted on tens if not hundreds of different Linux distributions. That wasted effort illustrates well how competition in the face of network externalities can impeded the toppling of an oligopoly. It's said that a camel is a horse designed by committee, and increasingly that's what Liunx for the desktop is looking like. There's no vision; but an ever evolving hodgepodge of elements that are almost but not quite ever ready for prime time. It takes as much effort to get a product from 90% to 99% as it does to get to 90%; so a lot of open-source products 'satisficing'; not perfect, but '(almost) good enough'.  

On top of this disorganization, costs are rising (partly due to the fragmentation of efforts) and time and money are running out. The open-source community has begun asking for monetary contributions (Wikipedia and Firefox this morning). I've made a few donations but after a while it becomes clear that this is a fairly futile exercise. No amount of money is going to lead to a product that just works.

It may take another fifteen years, or it may survive for a tool for a very few, but Linux will never compete in any meaningful way with Windows and OSX in the non-technical desktop market. Anyway, the battleground has moved away from desktops; Sony won the format wars the second time around (having lost to VHS) but Blue-ray's victory is phyric as content delivery moves to the internet. Microsoft (with Apple a good runner up) has won the desktop, just as the focus moves to tablets and phones: which means that innovation in this space will decline and what we have now will be with us for a long time.

So I'm considering moving back to Windows as my main platform. I'll still use Linux for coding, but for most of my day-to-day (reading, writing and emailing) not spending quite as much time on platform maintenance looks increasingly attractive.