I just submitted https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=105326 which contains a simple patch to try to improve the estimate for package installation time. My test seems to show that it is a good improvement (though still by no means perfect). I'm interested in any comments others might have. The basic idea is that instead of just using the install size to estimate time: finishTime = totalSize/sizeCompleted * elapsedTime we can also use the number of packages finishTime2 = numTotal/numComplete * elapsedTime and combine the two to get an estimate to show the user: realGuestimate = sqrt(finishTime * finishTime2) Elijah
