3 ms·
That may well be (and there are techniques believed to have helped significantly, like automated unit-tests and automatic memory management), but the outcome is
by pron 6y ago
That may well be (and there are techniques believed to have helped significantly, like automated unit-tests and automatic memory management), but the outcome is still worse than Brooks's prediction, which many rejected for being too pessimistic.
- AnimalMuppet 6y agoI'm not sure I agree with that. I suspect that we have enough libraries available today that I can write 1/10 of the code. That's a 10x speedup for my programming.
- nindalf 6y agoToday I can write an app that supports every language, not just English + (optionally) latin scripts. I can support every operating system, without worrying about platform quirks like fonts or rendering. I can deploy to it be available to every person on the planet within minutes, because I don't have to worry about managing my own hardware. I can harden my app in advance to security problems thanks to the knowledge that's out there. I have access to memory safe, highly efficient languages that didn't exist or weren't mainstream even 15 years ago. If you asked a competent programmer to do the same thing 30 years ago, they would take at least 10 times more effort than I put in, if they finish at all. Likely they'd give up on Linux/MacOS support, give up on Chinese, Japanese and Korean support, code it in 90s PHP and hope they didn't have any security problems. All of these are actual productivity improvements. Don't just measure time taken to complete a project. Measure actual work done and features shipped. We're doing more now, because we can.