• deng kun (8/8) Oct 09 2003 Hi, I'm not a real d programmar after all, but after reading this online...
• Justin Henzie (32/43) Oct 09 2003 Respectfully, I would have to counter your argument concerning the size

deng kun <deng_member pathlink.com> writes:

Hi, I'm not a real d programmar after all, but after reading this online manual
of D language and also some of the other web materials on performance
comparisons of d/c++/java, I really think d has a great potential in becoming a
substitute of c++. But the problem is it can't beat java if its language design
is still kept in a small circle, and the size of its runtime library is still
limited by the ideas of being "just sufficient for programmers to work on".

Anyhow my suggestion is D language should be made more popular by following
either the "GNU" way and/or Sun's standardization way.    

Justin Henzie <jhenzie mac.com> writes:

Respectfully, I would have to counter your argument concerning the size 
of the runtime library.

It is my sincere belief that the phobos library should remain light 
weighr, with an STL type collections module and perhaps a more 
sophisticated network module being stongly desired.

The other libraries that java programmers are oh so proud of should not 
be part of D but optional Libraries provided by the community or 
corporations.

Not only does this provide an opportunity for people to make some moolah 
from their efforts in D, something sun seem determined to prevent given 
the all inclusive strategy of java, but also allows developers to select 
appropriate libraries without necessarily bloating the size of their 
executable or library.

Java has a plethora of libraries available to it, how many of them do 
you regard as bullet proof.  java.io and java.net perhaps.  The 
collections are a poor substitute for STL but OK if you have never 
experienced anything better, its a bit like the windows argument really.

Moreover, D is a development language, it is not intended as a platform. 
and thank goodness for that, Providing a UI library as part of phobos 
would necessarily require compromises that would make it s poor 
alternative to native toolkite, witness the struggling, nee failing of 
SWING and AWT.


productivity enhancements (GC) but returns control of the computer to 
the developer.  D ROCKS, not as much as Objective-C but its faster <smile/>

D is appropriate for systems development, desktop development and server 
development, they are very different domains and require fundamentally 
different libraries, I am not sure you want to be carrying that baggage 
around with evey application you develop.

Regards,

Justin



deng kun wrote:

 Hi, I'm not a real d programmar after all, but after reading this online manual
 of D language and also some of the other web materials on performance
 comparisons of d/c++/java, I really think d has a great potential in becoming a
 substitute of c++. But the problem is it can't beat java if its language design
 is still kept in a small circle, and the size of its runtime library is still
 limited by the ideas of being "just sufficient for programmers to work on".
 
 Anyhow my suggestion is D language should be made more popular by following
 either the "GNU" way and/or Sun's standardization way.