[concurrency-interest] Reference/IdentityMap toString

Jason Mehrens jason_mehrens at hotmail.com
Wed Aug 25 14:26:43 EDT 2010


Tim,
 
>>"I don't accept the premise that AbstractMap.toString "suffers from some issues" when used as part of an implementation of an identity map. AbstractMap.toString is not broken or deficient;" 

I listed four failure cases and backed it up with examples (uniqueness was only one aspect).  Where is the counter argument that AbstractMap.toString is right for an identity context?  No one has defined why it is right, best, or only choice.
 
 
>>you are asking too much of it. You're trying to turn it into a serialization mechanism, something it is not meant for. 
I was hoping for creative insight on map theory but, I've only succeeded in stirring up creative insight on what I'm doing and why I'm doing it.


 
>>If you feel you really do need uniqueness of string representations for your identity maps, you can achieve it without imposing extra complications for those who don't. Several people have given examples of how you might accomplish this in practice.
This doesn't answer first question of my first post on this thread.  I'm not going to accept counter arguments of "why do that", "don't do that", "don't see the need", "use a workaround", or "don't mess with the JDK" when I'm asking about theory.
The points you raised on bijection were very helpful.
 
Thanks,
 
Jason

  		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.oswego.edu/pipermail/concurrency-interest/attachments/20100825/054c63e6/attachment.html>


More information about the Concurrency-interest mailing list