|
10 years ago, before I went to Iraq, I was a programmer at a start-up. We were trying to replace the expensive and proprietary ways that libraries index periodicals (I never said it was terribly exciting...) The system was based on the principles of open source software (Linux, Firefox, Open Office, etc.): you get to benefit from the work of others as long as you also donate your own work in the process. (For instance, you can get the insanely-cool, eyecandy-rich, beryl-powered, proactively secure GNU/Linux operating system for the cost of precisely nothing, other than the agreement that if you yourself make any improvements to it, you have to let other people use those improvements with attribution). This is also, essentially, the operating principle of Google: trust the users. Trust the community. They'll get it right more often than you will.
Bottom-up. Periphery-in. This was what was letting us compete with with the McGraw-Hills and Bell & Howells of the world. Not just compete, but win; we eventually got into like 75% of university libraries, and were one of the companies that ultimately morphed/merged into DOAJ.
I've taken some mean shots in the generational bloodbath we have going on here. I won't deny it. It's kind of a leftover from my fights in conference rooms 10 years ago, when we were struggling to wake business up to the tsunami that was coming. I still believe in open source. I still believe in the commons. Bottom-up. Periphery-in.
Back in those days, Linux was still a joke in server room. Once x86 hardware started to be server-capable, it was all Windows NT, all the time. And RISC platforms? Well, IBM, DEC, etc. had their own proprietary UNIX-workalike that only ran on their chips. But it ran great, right? And the userlands were almost compatible with POSIX. Linux was an untested, risky pipe-dream that naively assumed that people would actually contribute back.
It's 10 years later, and DEC is out of business (technically part of HP now, I think). HP, IBM, and Sun have pretty much given up on their proprietary OS's and use Linux as their method of selling their RISC platforms (and, damn, I wish a RISC system would really take off on the desktop...). Mainframe computing is dead, dead, dead and distributed systems are universally looked at as the way to go forward: Va Tech's outstanding supercomputer is a cluster of commodity Mac G3's. Google's clusters (I've seen the ones in Falls Church) are rack after rack of commodity hardware running a Linux variant. 10 of the root nameservers (I used to have a rack next to one; it had an armed guard) are on Linux.
The chorus of power-tie-wearing CTO's and CIO's hemmed and hawed and said it couldn't be done and would never work. They condescendingly thanked us for our time and showed us the door. But the few more perceptive among them kept an eye open, and saw we were doing it anyways.
My point is, I've been here before. I've heard the powers that be tell me I don't know what I'm up against and I'll never survive the real world. Our start-up went public for several million. Our technology helps researchers all over the world make new developments. Bottom-up. Periphery-in.
The side project of a Finnish grad student has, in 15 years, become the reference platform for all serious server deployments, because of the work of a committed network of people who believe in the common vision it represents. It's not only possible, it's been done.
|