Slashdot Mirror
Software Development Practices At Google
prostoalex writes "It's widely known that Google allows its engineers to spend 20% of paid work time on personal projects (that are nevertheless considered property of the company). But how does the practice actually work? Joe Beda provides a helpful insight in his blog, pointing out other interesting software development practices at Google. The code database is open for everyone, the snippets and pieces are documented and one is encouraged to re-use existing code. The intranet is transparent to the max and the company accepts the fact that there's more than way to accomplish something, so a better method is always welcomed. Interesting to note that just like Hawaiian shirt days in Office Space, the 20% per projects are "actively encouraged" - Joe suspects his review ratings might slip if he doesn't have one soon."
Here
"I'd rather be a lightning rod than a seismometer." -Ken Kesey
Actually google gives a very large (what that means I don't know) bonus to a person that made a very successful project. The problem is this generates a culture of people that get a project to beta, get the money and then drop development in order to get a new project into beta. Its one of the reasons why google has so many beta projects and finishes only a few of them.
"You can now flame me, I am full of love,"
Except California law states that if you do it on your own time and on your own equipment, you own it. Its not legal to request an employee to sign that away, and is not enforcable. Burden of proof is on the employer to prove the employee did it on work time, rather than the reverse. In this state those IP agreements are a scare tactic, nothing more.
I still have more fans than freaks. WTF is wrong with you people?
Parent poster may know this, but I wanted others to know: Fred Brooks is also one of the people who laid the foundation for modern computer architecture back in the late fifties/early sixties. The Mythical Man Month isn't clear about this, IIRC, but Fred Brooks was actually the lead _hardware_ architect on the IBM System/360 project, and then later the OS/360 manager.
This is costing somebody a funny point, but as somebody who did read the NKS book from cover to cover, including the copious notes, I can't let this popular but blatantly false rumour be repeated again unchallenged.
If Wolfram did anything that might be questionable it was largely cutting himself off from others who were working in the same areas for the decade he was predominantly focused on the research that went into the book. Yet it's obvious he wasn't expecting it to take more than ten years when he started and it obviously got to a point where getting it finished and out there became his dominant motivation.
Sure, he doesn't get everything right, and he makes some leaps that aren't all that convincing, but those faults are only a tiny fraction of the totality.
To me one of the most impressive things was the way he presented the background stories to all the key topics. Those I know a bit about were accurate enough that I can only presume the others were to.
The book is the way it is largely because he convinced himself that he was onto something significant that others had missed. However that something was certainly not, as neophites seem to assume, the general field of cellular automata, but rather his "Principle of Computational Equivalence"--an ambitious claim which he felt could only be presented in the context of substantial research.
-- Our systemic servants do not good masters make.
Except California law states that if you do it on your own time and on your own equipment, you own it. Its not legal to request an employee to sign that away, and is not enforcable. Burden of proof is on the employer to prove the employee did it on work time, rather than the reverse. In this state those IP agreements are a scare tactic, nothing more.
No, that's not at all true. The first exception in CA Labor Code Section 2870 says it doesn't apply to inventions which:
(Feel fre to Check it yourself.)
If you work for a large software company, that covers a *lot* of ground.