D benchmark code review

Joseph Rushton Wakeling joseph.wakeling at webdrake.net
Sat Dec 14 10:46:38 PST 2013


On 14/12/13 17:22, "Ola Fosheim Grøstad" 
<ola.fosheim.grostad+dlang at gmail.com>"@puremagic.com wrote:
> What makes me sit on the fence regarding D (I have used it actively a couple of
> years ago) is:
>
> 1. Not enough improvement on syntax (In some areas better than C++, in others
> worse. In regards to templates it is even worse, and C++ is kind of bad.)

Very curious about what you mean here -- I think most of us find that D's 
template syntax is much superior to C++, so I'd really like to understand what 
you find problematic.

> 2. No way to get rid of garbage-collection without making the language crippled.
> This is a show-stopper.

Out of curiosity, is this a show-stopper for theoretical reasons, or did you 
actually run into practical problems with the GC?  I ask because despite having 
fairly stiff performance requirements, I've not generally had any problems 
getting C/C++-like speed while using the GC.  Then again, I don't rely on any 
kind of real-time performance, only overall speed.

> 3. No high performing authoritative compiler suite. When the efforts are spread
> over 3 compilers I just don't expect any of them to improve to a state where it
> becomes excellent (like having excellent error-messages, analytic features,
> tight IDE-integration etc). It gives an impression of a lack of direction and
> leadership, and makes me feel like there is no hope of D ever to catch up. Other
> languages keep improving too…

I don't think things are actually as spread out as you think.  In reality, there 
is one compiler frontend that everyone collaborates on.  That frontend is ported 
to different backends, which each have different advantages in terms of speed 
and optimizations.  But most of the things you are concerned with seem to be 
things that depend only on the frontend.

> So obviously, it is not the semantics that makes me a lurking fence-sitter. The
> issues that makes me sit on the fence are certainly in areas that could be
> fixed, but I don't expect it will be. So I stick to the forums, for now… ;-)
> Though I do really wish you the best, and will certainly use/contribute to D
> when/if it resolves the issues listed above.

I think it might be worth giving things another spin.  If the last time you 
tried using D was a couple of years ago, you will find that things have changed 
a great deal in terms of performance, features and general quality.

And, if you're up for contributing, why not see if you can address some of the 
issues you raise?  For example, as you want to avoid the GC, one area that could 
use some input is going through Phobos working out what allocates when it 
doesn't need to, and fixing it.



More information about the Digitalmars-d mailing list