On Monday, 29 May 2017 at 01:36:31 UTC, Mike Parker wrote:
More broadly, I think what we need to be doing is teaching people that D's GC is not their grandfather's GC and that, unless they are doing something highly specialized, they probably don't need alternatives. The GC is fine by itself for a number of apps and, where it isn't, mixing in a bit of @nogc is probably better than cutting it out altogether.
One fun tutorial would be to integrate with a tedious C++ framework and let the GC take care of allocations in C++ code where speed doesn't matter.
Then write a C++ integration tutorial around it. That could be a selling point.
