On Wednesday, 23 September 2015 at 09:44:10 UTC, Chris wrote:
Yes, it's hard to beat experience. However, if a feature -
albeit useful - would break too much code, Walter takes a
conservative stance, else D would lose most of its clients. We
need a proper transition strategy.
C++ seems to take a guideline + guideline support library (GSL)
in order to "softly deprecate" the old feature set without
actually deprecating it. I am not sure how well that goes, but
GSL looks good to me.
When a language reach a certain complexity it probably is a
mistake to add more features, new features have to be followed by
instruction of how to make good use of them. In C++ there are too
many ways to do the same thing, which is negative for code
legibility.
His talk is online:
https://youtu.be/1OEu9C51K2A
The question is, do certain issues really keep you from using
the language successfully and why? And is it an issue for a
substantial part of other users?
Another question is: what kind of competing solutions are
emerging. Herb Sutter seems to have focused his cppcon talk on
Rust style memory management in C++. The adoption of Rust does
force the C++ designers to switch gears and hopefully the
competition will create a push for better solutions.
That applies to D too, I think.
Rust also aims for commercial level stability for Rust 1.4. (I
don't think they will be able to, but let's see.) This is also a
threat for C++.
