> From: Stuart Ballard [mailto:[EMAIL PROTECTED]]
>
> <snip>
>
> For a spec that is supposed to be stable and implementable, beta4 has
a
> long way to go - I found another significant spec bug in the latest
> changes to java.util.List also, which will make some common methods in
> spec-conforming classes O(n) when a better designed spec could make
them
> O(1).
>
> <snip>
I remember that they did specify the O() of many of the classes, mainly
to highlight the tradeoffs you would take if you used the class. If
your implementation makes the big-O of any of your operations greater
than Sun's, or requires a great deal more space, then you are just using
a different tradeoff. It's best in that case for you to just create a
new gnu.java.util.BetterList class that uses the new tradeoffs.
I don't know if that's the case. Perhaps you do just know a better way
to do it and still conform to the spec.
--John
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
