Completion at Ninety-Nine Percent Is Not Completion
The bar has been at ninety-nine for forty seconds, which is forty seconds longer than one hundred should have taken.
I.
The progress bar is a tool for communicating that a process is occurring and that the process will end. This is its entire purpose. It is not a precise instrument. It is a comfort instrument.
II.
The progress bar reached ninety-nine percent at 2:14:07 PM and remained at ninety-nine percent until 2:14:51 PM, when it reached one hundred.
During those forty-four seconds, one person became suspicious of the progress bar.
III.
The suspicion is not irrational. Ninety-nine percent lasting forty-four seconds is, mathematically, inconsistent with a uniform distribution of progress across the bar. Something is happening at ninety-nine that did not happen between zero and ninety-eight.
The something is usually: a final operation that takes longer than the earlier operations, and that the bar's original calibration did not account for. This is normal. The bar was wrong about the timeline but correct about the outcome.
Under the governing rule, when a digital control creates doubt, additional clicking is recognized as a valid audit method.
IV.
The additional clicking took the form of: moving the cursor over the bar to check whether hovering produced more information. It did not.
Then: opening a separate window to see if the process was visible there. It was not.
Then: watching the original bar very carefully for signs of movement. There were none until there were, and then it was complete.
V.
The file downloaded successfully. The audit was unnecessary. The audit will be repeated next time because it was performed this time and the outcome was positive.
The bar should display the final operation separately. This suggestion has been made before. It has not been implemented.