Re:Go’s alias proposal and all my concerns of Google control...
by iainmerrick on Monday November 07, 2016 @10:52AM
(#12891956)
This has an interesting interaction with another of Go's quirks.
As I read this (and it seems like the author has an axe to grind so maybe they aren't explaining the motivation correctly) the idea is that when you want to rename A to B, you can make an A -> B alias to ease the transition.
However, people calling your code and using A won't know that they need to update their code unless there's some kind of warning. But Go doesn't have compiler warnings.
So I think the author is right, this would be better implemented as a compiler error plus an automatic refactoring tool. That approach would leverage one of Go's strengths -- it's a pretty regular language and already has good refactoring tools.
As I read this (and it seems like the author has an axe to grind so maybe they aren't explaining the motivation correctly) the idea is that when you want to rename A to B, you can make an A -> B alias to ease the transition.
However, people calling your code and using A won't know that they need to update their code unless there's some kind of warning. But Go doesn't have compiler warnings.
So I think the author is right, this would be better implemented as a compiler error plus an automatic refactoring tool. That approach would leverage one of Go's strengths -- it's a pretty regular language and already has good refactoring tools.