Stateless comp question

[wh72 at cornell.edu (William Hatch)]

on 7/9/03 4:10 PM, Chuck Hill at chill@xxxxxxxxxxxxxxxxxx wrote:

> I'll first comment that you are probably wasting your time.  The
> differences in any of these implementations is unlikely to make a
> significant difference unless  you have many bindings or make
> extensive use of these components.  The real savings with stateless
> components is avoiding the overhead of repeated object construction,
> destruction and garbage collection.  I'd want an optimizer to show
> there was a material gain to be made before I spent too much time on this.
> 
Well, this particular component is wrapped in a long repetition, so I
thought it might be better memory wise.  But I think you were right, I
certainly didn't notice any performance gains.  I added the synch method to
takeValues, invoke, and append and it all worked.  It doesn't seem logical
that by making it stateless, you have to synch more than just making it
non-synching; it seems counter productive, but I know nothing.  Anyways,
thanks for the help.


-- 
Bill