Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

Re: [News] More Buggy Progress Bars from MS

On 2006-11-11, Erik Funkenbusch <erik@xxxxxxxxxxxxxxxxxxxxxx> posted something concerning:
> On Sat, 11 Nov 2006 14:25:49 GMT, Sinister Midget wrote:
>
>> I used to have to copy several megabyte from a networked drive to the
>> local machine and back to another networked drive while running XP. It
>> took no more than a handful of minutes from drive to machine, and the
>> progress bar wasn't too awful inaccurate (jumping between displaying 10
>> minutes, 1 minute, 5 minutes, etc left randomly). The operation from
>> machine to network took about an hour. The progress counter for the
>> second leg would gradually rise to 10 hours, drop to an hour, drop to
>> 45 minutes, go to 8 hours, drop to 3 hours, go to 14 hours, drop to 2
>> hours, etc.
>
> It's not intended to be accurate.  It's an estimate.  Over network
> connections, in particular, Windows doesn't scan all files and
> subdirectories first, so it estimates the time based some average number of
> files per folder at a certain size, but if this turns out to be different
> when it actually gets to a subfolder, it adjusts its calculations.
>
> Many operations, especially those involving compression, are hard to
> accurately estimate.

Thank you for supporting my point. It can't be estimated accurately, so
the attempt to accurately protray the length of operation is less
valuable than worthless.

>> Progress bars are one thing. Trying to time the activities they're
>> supposed to be measuring isn't among the brightest ideas anybody has
>> come up with since they seemingly can't account for all of the
>> variables that affect what the operations are doing. So it's probably
>> best to leave things as a measure of overall size and completion
>> instead of overall time.
>
> Despite the inaccuracies, I think time estimation is helpful.  In most
> cases, it will give you an idea if it's going to take a few minutes or
> several hours.  

Except when it's so grossly inaccurate as to be completely worthless.

My example of stuffit in my previous post is a good example. It was
decompressing a file locally. It was counting the numbers of files
being extracted. I saw no hesitation on any of the extractions that
would cause major shifts in the expected time to completing. Yet it
counted continually upward until it was well past sanity, and completed
just a few seconds later. In fact, the time to completion adjusted
upward at a rate of 2 or more minutes to every file extracted. In never
decreased until the very end. What good was the progress counter?

-- 
Windows? WINDOWS?!? Hahahahahahehehe.....

[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index