I use Time logging against Original estimate to give me an hourly Burndown chart, so the value is earned immediatley (as opposed to waiting untils it's 'Done')
My discussion point revolves around the impact logging time has on the remaining estimate, and if that estimate is too small what are peoples psrference as to how to reflect that underestimating?
For example if you have a 4 day estimate, and a resource has spent that already, but has 2 days of actual work left to do? Do you create a seperate task for the 2 days, and flag it accordingly so you can assess how many underestimates have been made at the end of a sprint?
Or do you up the estimate? What are the advantages of doing this?
On the flip-side of this if a task is finished early and 2 days are left on the estimate, how is this earned? (this may be a simple one as I think you can configure for the full value to be earned when its moved to 'Done')