Final by default?

[Walter Bright]

On 3/13/2014 8:56 PM, Manu wrote:
> It does nothing to prevent the library case,

I addressed that in another post here, i.e. I'm very skeptical of the notion 
that a large code base built with no attempt at writing fast code is going to 
get fast with merely this change. I don't buy that high performance code comes 
by accident. If you wish to debate this, though, please reply to my other post 
on the topic, which expounds with more detail.


 > or the don't-early-optimise case of implying breaking changes in the future.

Optimizing code often involves breaking the interface. For example, the 
existence or absence of a destructor can have dramatic differences in the code 
generated. (Even very experienced C++ coders are often unaware of this.)