This is a good list, and a great starting point for a PhD. thesis. It's the first really useful post I've seen in this thread. Too often we underestimate the "non-technology" factors (like supportive management, and a market for the end product) that go in to creating good software.
I'd like to add one more item: Making the informed decisions about what to optimize. Every technology optimized for *something* -- execution speed, compiled/downloaded code size, source code elegance, system flexibility, portability, developer efficiency, shiny user interface goodness... the list goes on forever.
I've seen too many projects created by otherwise smart developers (with supportive management) that have failed because they optimize for the wrong things -- for execution speed instead of portability, or for portability instead of developer efficiency.
To cannibalize Sun Tzu: Every bit of software is meant to operate within a specific domain. If we don't understand that domain then the software has little chance of being successful.
Best of luck figuring out how to MEASURE these things.
Slashdot Mirror
This is a good list, and a great starting point for a PhD. thesis. It's the first really useful post I've seen in this thread. Too often we underestimate the "non-technology" factors (like supportive management, and a market for the end product) that go in to creating good software. I'd like to add one more item: Making the informed decisions about what to optimize. Every technology optimized for *something* -- execution speed, compiled/downloaded code size, source code elegance, system flexibility, portability, developer efficiency, shiny user interface goodness... the list goes on forever. I've seen too many projects created by otherwise smart developers (with supportive management) that have failed because they optimize for the wrong things -- for execution speed instead of portability, or for portability instead of developer efficiency. To cannibalize Sun Tzu: Every bit of software is meant to operate within a specific domain. If we don't understand that domain then the software has little chance of being successful. Best of luck figuring out how to MEASURE these things.