[OT] Modules dropped out of C++17

[Nick Sabalausky via Digitalmars-d]

On 06/09/2015 04:58 AM, Shachar Shemesh wrote:
>
> I would have a considerably less problem with UFCS had it been limited
> only to explicit declaration.
>


That's how it works in C#. That was actually my first introduction with 
extension methods, and then when UFCS was added to D, I used to be 
pretty vocally opposed to it being implicit. But I learned to live with 
it, and then I eventually I realized it was basically *never* causing me 
any real trouble, and now I rather like not having the cognitive load of 
"would this be better as an extension method or not?" every time I 
define a function. Plus, I don't have to deal with the issue of calling 
a func where *my* code could have been much cleaner had the func's 
author made it an extension method, but they didn't, so I'm stuck 
dealing with it or going to the bother and mess of a wrapper.

I see it as a stylistic issue, now. Let the caller use whatever works 
best for their own code.

Sure, it makes it possible for the caller the use UFCS inappropriately 
and obfuscate their code in goofy ways, but in practice this is almost 
never a real issue. And besides, pretty much *any* feature can be abused 
and obfuscate code. Ex: Nothing prevents me from naming a file-loading 
function "save" (BTW, I've actually seen that in commercial production 
code - wasted half a week figuring that one out), but being able to 
choose arbitrary identifier names is still a good thing.