On Thursday, September 22, 2011 11:15:23 Regan Heath wrote: > as results speak louder than proposals I have found.
It's often easy to come up with ideas. The hard part is implementing them, and in many cases, they need to be implemented to be shown that they're viable and to iron them out. Also, there are only so many people working on implementing stuff for dmd, druntime, and Phobos, and they only have so much time - not to mention their own interests - so simply suggesting a good idea is often not enough to actually get it implemented, even if it _is_ a good idea. So, while good ideas are most certainly welcome, and discussing ideas is often quite advantageous, ideas will get a lot farther, faster if the folks suggesting them actually implement them. This is particularly true for compiler features. Walter doesn't spend a lot of time implementing enhancement requests at this point. He's working more on core implementation issues and bugs. So, even if a feature request is solid, backwards compatible, and generally considered to be desirable, it's a _lot_ more likely to make it into dmd if someone else takes the time to do it. Fortunately, the level of contributions to dmd has been quite good of late and the number of people who have some understanding of dmd's implementation is increasing, so dmd has been improved at a considerably increased pace. So, in any case, ideas are great and welcome, but it's often necessary to volunteer your own time and effort to actually get your ideas implemented. - Jonathan M Davis
