Comment by dhimes
13 years ago
The concept of "premature optimization" also has another connotation in product development: Don't waste too much time making that product or feature optimized until you are convinced you can actually sell it. It's not that optimization is bad, but optimization before market trial (premature) can result in you spending precious time working hard on the wrong thing.
Optimizing the right thing is good, but figure out what that thing is first.
In a publicly traded company, such as Microsoft, that doesn't take long. It is shareholder value. It is not Optimization "for its own sake" promoted in the article. I savvy developer made need to run business tests against their software.
> It is shareholder value.
But in what timeframe? The article seems to imply they are thinking only about short term value and ignoring medium and long term.
The article implies that line kernel workers interested in optimization for the sake of optimization are ideally situated to judge what matters in the long and medium term, and that Bill Gates and Steve Ballmer, despite being the two largest share-holders and more than three decades of experience running a software company (apiece) are incompetent to recognize long and medium term strategies.
The fundamental assumption of the article - that the Windows kernel should strive toward the Linux kernel management model is patently absurd.
FOSS is free as in "free to break software that depends on your code."
6 replies →