🚀 go-pugleaf

In article <aS9B5.60842$c5.169252@newsfeeds.bigpond.com>, Paul Murray
<pmurray@nospam.bigpond.com> writes
>"Dave Bird" <dave@xemu.demon.co.uk> wrote in message
>news:RfGmCUBpl405EwfS@xemu.demon.co.uk...
>
>>  Java deals with the issue of classes better and more cleanly.  Java is
>>  definitely the nicer one to program in, though you may be less pleased
>>  with the ultimate performance of the result on your particular machine.
>
>OTOH, gigaHz processor machines with half a gig of memory in them are
>beginning to make performance irellevant. Java comes with a standard set of
>GUI widgets, but they are buggy, goddamnit. The bugs are mainly when you run
>it under windows, because of all those undocumented windows API things that
>M$ don't tell anyone about.

 I'm unconvinced that speed issues will ever "go away", because
 people just expect that more and more elaborate things can be
 done at blistering machine speed.   

 I think you will always have to live with Java been that much
 less perfectly tailored to the individual machine, that much
 less fast performing, that much more lacking in fancy features 
 when you finally need them, and so forth.  These are just
 the price of its portability, security, and ease of use.


 If you are saying that "over-all the language writers will get
 better at their craft", then this would be measured in the extent
 to which each language overcame its limitations.  If, for instance,
 Java offered a cleaner transition to fully compiled code, got the
 bugs out of its widgets, got a better clue about operating with
 the graphics on individual machines, etc.  

 Equally C++ could offer better handling of classes, offer the
 option of enforcing a single inheritance model, deprecate some
 of its more annoying features, offer some help on memory allocation
 and de-allocation.  It would be difficult to offer a garbage 
 collector model, as each instance of a pointer to dynamically
 allocated storage would have to become a handle.  I suppose it
 is possible to have temporary copies of the handle provided the
 handle stays on the same allocated range for the duration.


 There are, annoyingly, steps forward which neither Java nor
 its model C++ have taken.  A unified postfix notation for
 allocating and accessing store [  fpi:()^int;  pfi:^()int;
 rJoe^.m14 = 0;  and so forth].   Remove the command as a
 separate idea to expression.  Better handling of arrays. Etc.




-- 
   ^-^-^-@@-^-;-^   http://www.xemu.demon.co.uk/
        (..)__u     news:alt.smoking.mooses