on 1-09-23 22.02, Dain Sundstrom at [EMAIL PROTECTED] wrote:
> The blueprint
> (http://java.sun.com/j2ee/blueprints/design_patterns/page_by_page_iterator/i
> ndex.html) applies to client access of the result set. The problem is the
> blueprint assumes that then entire result set of the query can fit into
> server memory.
The implementation do not assume ! ... ? ... :
-specifies that ?
-get count of all --list 10-50 first (pk) for client view
-on way back (from response sent) hash all pk in cache
-wait for new input from user
-use (*) built index cache for next
-and have a smart cache for Your app ...
*idle time = MEANWHILE prepare for users next move
Otherwise - I gave to much cred to the pattern -and are sceptical to why
institute it -at all ? ...
Love
/peter_f
_______________________________________________
Jboss-development mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-development
