Do we need to indicate that an activity that ran very quickly didn't error-out immediately?

Question

We have a system that has a few maintenance activities that run very quickly (less than a second). Problem is, when users look at the log file for the system, seeing an activity that ran for 0 seconds could look like an activity that failed to launch entirely. Mind you, when an error does occur, the log will clearly show that an error occurred.

Is there reason to believe that users unfamiliar with the system would assume that activities running for less than a second would mean the activity never completed successfully?

How the log file currently reads:

Started at Aug 28, 2012 05:28:15 UTC
Stopped at Aug 28, 2012 05:28:15 UTC (0 seconds)

Answer 1

If errors are clearly indicated, then users can assume that a task that is not marked as an error did complete successfully. However, if most users may have never seen the record for an errored task, then this confusion could occur. Here are my two suggestions, and in my opinion, you should utilize both, but one should be be sufficient:

• Change 0 seconds to 0.1 seconds. If the task ran, it must have taken some amount of time, so 0 seconds is not accurate. Since tasks complete with that kind of precision, your results should show decimal places, but if you want to or need to round it to the nearest whole numbers, then mark this time as <1 second.
• Change the term "Stopped" to "Completed". A task could have stopped suddenly if it errored so I think that completed conveys that it ran successfully and did not stop prematurely. If you can not change this term due to the errored tasks (assuming it reads something like "Stopped at Aug 28, 2012 05:28:15 UTC (0 seconds) [ERROR]"), then maybe you could add a similar indication like [SUCCESS] or [COMPLETED] so there is no confusion about the cause of the stop.

Answer 2

I'd recommend running a quick test to actually prove if this is the case. A five second test should be enough, you don't need to ask a lot of users, a small-medium sample will probably give you a good idea of what's going on. I would probably ask something in the lines of "do you think this process has been completed?". You can find usability testing tools online, some are paid and some are free to use (I quite like FiveSecondTest).

An alternative to consider would be to have a default running time log of, for example, 0.1 seconds, just to show something other than 0.