[ http://jira.jboss.com/jira/browse/JBPM-492?page=comments#action_12348753 ] David Lloyd commented on JBPM-492: ---------------------------------- The current solution is that I'm just fetching the first N pages of entries (right now it's 100). The database is only hit when the search criteria change. This way, the result set is finite, but a database hit is not required for each page change (which means that moving from page 1 to page 2 back to page 1 always gives you the same result, until you apply new search criteria). The logic behind the first N pages is that, as a user, more than 100 pages is not useful to page through. If you're looking for a task, you'll use the search criteria to narrow down the result set. This would also allow us to add sorting, because the results can be sorted without hitting the database. The only problem is that sorting can lead to misleading results when you've fetched the maximum number of rows; therefore in this situation, the user should be warned that their result set is too large and they must narrow down their search before sorting will function properly. As an optimization, when applying search criteria, the sort criteria could be applied to the query before fetching the results to prevent a re-sort on the client side (for many sorts, there is a possiblity that the database could use an index to optimize the sort). Of course since the server might have slightly different sort semantics than the client, the client would re-apply the sort just to be save, but it should be able to run the sort in O(n) time (sorting an already sorted set). -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira