For me, my C code is very micro-efficient so it generally runs faster than Go, but it takes me up to 4 to 5 times longer to write things in C than it does to write things in Go. If the resulting Go code is not fast enough, I have three solutions:
1. Run the profiler to find the bottleneck, sometimes I can fix it. 2. Complain about the performance on GoNuts and when Go updates, my code usually ends up running faster. The devs are working hard so we don't have too. 3. Give the code to my eldest son Daniel, he is a much better Go programmer than I am and he can optimize the hell out of any bottlenecks. When I wrote my first C program in 1988 or 1989, it was a true color video card driver for IBM and micro-efficiency was very important. Before that I used Forth, MacroASM, UniComal on Fortan77 Runtime. Today, it is rarely important and the profiler can be used as a tool for cost justification for any efforts to speed up a program, but mostly, it is not much of an issue anymore for me. My concern is time to complete the project and cost to complete the project. If it is a small poorly defined project, a talented PHP programmer can interface with the stakeholder and complete the project almost as fast as a Go programmer for much less money because here in Arkansas PHP programmers are being train in the local colleges and are relatively cheap while experience Go programmers want three times as much money for their time. For a larger well defined project Go is just much easier to partition out the work amongst a team working parallel. And for me, personally, Go is just much easier to be productive. -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
