The progress bar illusion | Hacker News Reader

The progress bar illusion

newscientist.com

125 pointsdchest34 comments

If a download or something seems to be taking too long, I'll actually put the point of my mouse cursor at the edge of the progress bar to see whether it's stuck or just moving really really slowly. I really don't like these illusions.

Apple has been using the "left moving ripple" progress bar for a while, which makes it look like the progress bar is moving, even when it's not

I wonder if it's a coincidence, or Apple knew what they're doing.

I think it's safe to assume that they knew what they were doing. Apple sweats the details, like making a sleep indicator light that pulses to the speed of a heartbeat.

http://gigaom.com/apple/the-pursuit-of-perfection-hidden-gem...

The pulse is much much slower than a heartbeat. More like breathing.

Yes, like breathing. http://www.patentgenius.com/patent/6658577.html

The moral of this story: Apple pays more attention to details than gigaom.com.

Ah yes, that's the patent I was looking for, upvotes all around :). I mistakenly googled for 'heartbeat' instead of 'breath', for which only the GigaOM article turned up.

"The average respiratory rate for adults is 12-20 breaths per minute, which is the rate that the sleep-indicator light fades in and out on most Apple laptops. Older models such as the Macintosh PowerBook, however, use a blinking LED indicator, with discrete pulses in one-second intervals.

The other day, I noticed that my friend’s Dell laptop had a similar feature but with a shorter fade-in-fade-out period. Its rate was around 40 blinks per second, or the average respiratory rate for adults during strenuous exercise not very indicative of something in sleep-mode."

http://floodmagazine.com/2010/10/14/apples-attention-to-deta...

That does not make the insanely bright white light any less annoying when you are trying to sleep, though. Quite the opposite, in my experience.

Bet yours is a couple years old. They really reduced the intensity when they switched to unibody enclosures.

Anecdote: My first MBP was a graduation present from my parents just before I graduated HS. For some unmemorable reason they “grounded” me from it for a day or two (during finals, maybe?) - but in its sleep state, it was keeping them from sleeping, themselves, with its bright, pulsing LED as it sat on Mom's dresser. I had to laugh just a little at what felt, to me, like Apple-sponsored karmic revenge :)

The current MacBook Air doesn't have a sleep light - the next round of MBPs might not have one either. The idea it to close the laptop when you are not using it. Sleep is the default state, so no need for a status indicator. Just like the iPhone/iPad/iPod.

There's a big difference between the MBP and the Air/iPhone/iPad/iPod though -- solid-state drives. The sleep indicator tells you when it's safe to pick your computer up.

I think it's pretty likely that the next round of MBPs will come with SSDs standard.

Sleep is the default state, so no need for a status indicator.

Wrong!

Sleep may be the default state but I definitely want to see whether it's really in that state - or in some other.

I hate this kind of wishful thinking from hardware designers.

It's the same mindset that removed the status LED from the Samsung Galaxy S. A status LED that exists on almost every other android handset and that is very useful for indicating missed calls, low battery and many other things (it's freely programmable).

Congrats Samsung and Apple, you saved on a $0.20 part - but in the wrong place.

Agreed. But, AMOLED gives some nice opportunity here. Try NoLED:

http://forum.xda-developers.com/showthread.php?t=730692

NoLED is a neat hack, but not a substitute for a built-in, reliable notification mechanism.

Nah, they're just doing it to be cool. Nothing in a quiescent, stable state (ie not requiring attention) should be animated. Animation - even a pulse - pulls your attention. Yes, you can /learn/ to ignore an animation, but you shouldn't have to do that in the first place.

In contrast with Microsoft, which switched from standard progress bars to "right moving ripple" progress bars in Vista. Which, as the article states, makes it look like it's moving slower!

Both still look like they're still moving when they're stopped though, which is the real point of them. Tricking the user into believing the task will be completed before it really will be isn't the goal - otherwise they'd intentionally underestimate completion times when they're displayed.

I think that actually is what they are designed to do, make it look like it is going to be completed sooner. If you just underestimate completion times, the you'll have a 100% complete bar while still making the user wait - that will make it look broken and slow, the opposite effect. Better to overestimate, then the task will be completed "early" because it was so "fast".

This is exactly how people perceive execution time. If a progress bar goes to completion and then hangs users will perceive that operation as taking longer than if a progress bar goes to, say, 70% and then jumps to completion, even if the total time of the operation is the same.

It makes sense though, people are the most impatient when they've been waiting the longest so being surprised that the operation completed faster than they were led to think it would is more welcome than having to wait longer than they thought they'd have to.

Vista SP1 changed the file copy algorithm to cache files, so that it would appear faster than it is (and other reasons):

http://www.codinghorror.com/blog/2008/03/actual-performance-...

Also, the progress bar that they used to have for starting up OSX (removed in Leopard I think?) had no actual relation to anything the machine was doing, it was just timed to move across at a particular rate based on the average startup speed of your computer.

http://daringfireball.net/misc/2005/04/tiger_details#waiting...

A completely fake progress bar--it was just something for you to look at.

That progress bar may not be tied to actual events, but it does give you a reasonably accurate view of the time remaining until the login window will appear-- probably more accurate than if the programmer guesses how much each startup phase is worth and the bar ends up moving erratically.

Ehm, I beg to differ. Booting off a regular drive (non-SSD) can produce pretty variable times, depending on if you just installed something, fragmented the drive with some large files, etc. I remember many startups on my old PowerBook G4 running Tiger where the bar would reach the end and keep pulsing for another twenty seconds, much to my confusion/amusement.

The biggest illusion of progress bars is when they don't accurately represent the time left for something to complete. Like when you wait 7,5 minutes for the bar to reach 75% and then it suddenly jumps to 100%, instead of taking another 2,5 minutes to complete (that is frequent on installation processes).

Of course, when it comes to download progress bars it's not possible to make it accurate without making it bigger as the download speed drops and making it infinite/disable it when the speed is 0KB/s. But in other offline tasks the progress bar is frequently useless to capture time remaining for the completion of something.

That's probably why a time remaining label is added to them.

Firefox is worse; it often immediately goes to about 75%, then just sits there, then jumps to the end; it is the most useless progress bar I have seen yet (and I have been using computers since Windows 3.1).

A small file could cause this. The browser starts to download the file to a temporary directory, while the user is choosing the final location, so when the progress bar appears, most of the file is already there, then the rest of the file arrives at the same time that the system is creating the new file and copying the contents from the temporary file to the location chosen by the user. Disk file operations can block for an instant. When the bar is to be updated again, the download is complete, so it simply vanishes.

Rethinking the Progress Bar: http://www.chrisharrison.net/projects/progressbars/index.htm...

A paper that investigates how non-linear progress rates also affect perceived completion time.

The brain is a tricky thing. Even seeing this video I'm sure that the "left moving ripple" progress bar still seems faster to us.

Can something like the be applied to the loading animations in webapps? If the circle pulses or slows then speeds up randomly will it seem like the app is working faster or slower for the user?

Despite the interesting content, I'm compelled to ask: does anyone actually watch progress bars?

I do. Of course, for long file downloads. Some programs do it right. But, as soon as I detect that certain progress bar is unreliable, I lose "interest" :-)

This reminds me of a old story...

The residents in an apartment block kept complaining that the elevator was too slow. The building's owners got several quotes to upgrade the elevator but they were all too expensive and ran into many thousands of dollars.

Eventually they were approached by a old man who lived on the ground floor. He said he could solve the problem for just a few hundred dollars. In desperation they agreed to give it a try and a few days later all the complaints stopped.

His solution - simply install mirrors next to the elevators. It was enough of a distraction that people didn't mind the wait.

I'm sure there are many more entertaining ways to distract than a simple ripple.

Interesting illusion, but I wonder what is the desirable effect here.

I can imagine it being argued both ways:

1. the bar should be animated so that it looks faster than it really is.

2. the bar should be animated in the reverse way, so that download is actually faster than it seems from the bar, thereby giving a positive surprise to the user.

How about the illusion of aging 11% faster due to the spinning rainbow ball that Apple has also been using?

Reply on news.ycombinator.com