Um, kids, Joel needs employees. So don't go to MS or Google, please?
Seriously, Joel? When was the last *important* thing you've built that didn't rely on some knowledge that came out of one of those two huge companies? The nice thing about being an architecture astronaut, is that every once in a while, you solve a real problem while trying to solve a different one. If MS and gogole didn't take on al these huge failing projects, how would the world learn about things like:
- Making huge teams scalable (or how *not* to)
- Creating very scalable architectures
- Identity
- WS Standards and Security
- Parallelism, synchronization
- Next generation languages and tools
- better tooling for existing languages
- Domain specific languages for solving complex and small problems
- lots more "little" solutions to many problems
People who work there don't work on hopeless projects, they go there to solve huge technological challenges that advance our software industry directly or indirectly by letting us know what the limits are in given contexts:
- A team of X size becomes a big problem if you manage it wrong (see "Vista Start menu Fiasco")
- Agile can work on larger teams (as long as you split them to smaller teams!) while people do other things as well
- Integrating winforms in the browser can lead to really bad things, but alternatives will be found
- How do you support X millions of transactions per minute?