Andrei Alexandrescu:

> I'm announcing this here because inevitably D will be mentioned during 
> the panel with Bjarne Stroustrup, Herb Sutter, Hans Boehm, and myself.

And also because several of the design ideas of C++ programs are useful in D 
code too :-)

---------------------------

About Bjarne Stroustrup's talk "C++11 Style":

I have appreciated (slide 22, at about 26.22) the stress on SI units, to use 
the type system and the new C++11 feature to some avoid some bugs with zero 
run-time cost. D requires a less natural syntax for them.

I have also appreciated the very simple but useful part about type rich 
programming, with typeful interfaces 
(http://en.wikipedia.org/wiki/Typeful_programming ). 

Another thing I've appreciated about Bjarne talk is that I'm able to understand 
it at about 1.3X speed (with just few pauses to digest certain slides).

The example of Vector Vs List (slide 51) is too much artificial, but the answer 
(in the invisible graph) is interesting. Lists are kind of dead noways.

The slide 64 "Low-level != Efficient": in D we'll write just:
sum(v)
Instead of the currently used:
reduce!{a + b}(0, v)
Because sum() is a very common need.

Overall Bjarne Stroustrup's talk was very good, and I agree with everything he 
has said and with his choice of topics to talk about :-)

---------------------------

About Stephan T. Lavavej's talk "STL11 - Magic && Secrets":

The "Computers are fast" slide at 20.31 is nice, it shows vector.emplace_back 
for a pair.

Seeing this talk at 1.4X is a bit tiring (I can't go past 1.6X because my video 
player is dumb, it's compressing audio in a uniform way, without taking into 
account the characteristics of human voice).

Regarding slide 25 (51.54), in D I'd like to use "const" and "const ref" in 
foreach loops, just like "ref":


struct S { int x; }

void main() {
    S[] array = [S(1), S(2)];

    foreach (ref s; array) {} // OK
    foreach (const ref s; array) {} // error
    foreach (const s; array) {} // error
}

Bye,
bearophile

Reply via email to