Steven Schveighoffer wrote:
On Sat, 06 Mar 2010 19:41:57 -0500, BLS <[email protected]> wrote:
..and frankly said I wonder .. no offense.. why you are the (only) one
who is creating and designing the collection library ?
finally :
from a pragmatic view > why create from scratch instead of reusing
Stevens data-structures..
We have had private discussions, and we have fundamental differences of
opinion for what belongs in containers. It may be that through trial
and error and constructive discussion we come up with the same
conclusions, in which case I'm sure we will combine forces (recently, I
think we both moved our points of view closer, but not quite there
yet).
I concur with Steve's characterization of the situation. My interface
definitions are very crisp and my cost function looks more like a step
function, so I'm difficult to negotiate with. I'm running a mile away
faster than you can say "Here, there's this Container.contains() method,
which..."
But Steve suggested that he could volunteer, with credit, some of his
implementation code to my interfaces, even though he has disagreements
regarding the exact definitions of those interfaces.
However, I will not stop maintaining dcollections as long as
Phobos' collection solution does not implement the features I think are
important. I am in the process of porting/improving dcollections to D2,
and when it is done, I think it will have some really useful features in
it.
That's great. This kind of competition can only be best for everyone.
Andrei
