Slashdot Mirror
More iTunes Math
markmcb writes "OmniNerd once again digs into the math behind iTunes. The 5-star ratings in iTunes are actually stored in a variable allowing values from 0 to 100, with 20 = 1-star, 40 = 2-star, etc. Known to few, if you set a rating to 30 (manually), it will show up as 1 1/2 stars in the program's GUI. Matt Schinckel provides interesting statistical evidence showing that not only do non-whole, non-half values (e.g., 47) not increase the amount a song will play, but neither do the half-star ratings (with the exception of .5 star)."
Do values outside of the range 0 to 100 cause the program to crash, for instance? Or does it just show zero or five stars, depending on which end of the range is being exceeded?
Cyric Zndovzny at your service.
Too bad itunes sucks.
So it seems they developed their algorithm to only deal with the values 0, 20, 40, 60, 80, and 100. Is it really a surprise that it doesn't take into account other values?
Cyric Zndovzny at your service.
Here's the Slashdot post to OmniNerd's first iTunes article: http://apple.slashdot.org/article.pl?sid=05/08/28/ 0616225
Direct link to the first article:
http://www.omninerd.com/2005/08/25/articles/34
Has anyone been able to replicate this behavior with the Windows version of iTunes? Is it built upon the same core as the Mac OS version, or does Apple have a separate implementation that may behave differently?
Cyric Zndovzny at your service.
This shows the beauty of scripting systems like AppleScript. They allow you a level of interaction with a system that is hardly seen on Windows, and only vaguely available on UNIX.
Cyric Zndovzny at your service.
I wouldn't necessarily consider it a bug outright. First of all, it doesn't appear to crash iTunes. But beyond that, it's an action that is outside of the normal usage scope. It requires one to use AppleScript, at least on Mac OS X, to use the non-standard values.
/dev/kmem? I wouldn't say so.
One might say that quality software should properly deal with unusual values, and that is quite true. Likewise, it appears to be the case in this particular situation. The rating system still works, even with the unexpected data. Indeed, could the Linux kernel really be considered buggy if somebody were to start entering random values into kernel memory via
Cyric Zndovzny at your service.
Linux is STILL for fags.
Itunes. But I'll tell you what I don't hate, The spicy italian sandwich from Subway. 2 layers of salami and Pepperoni topped off with 4 slices of american cheese. Put a little mayo on there, some lettuce, tomatoes, and black olives, and you've almost got yourself a sandwich. Next, have them cover the sandwich in salt and pepper, then spray some oil and vinegar over the top, wrap it up and there's dinner. I think we can all give thanks for a sandwich that good. mmmmm mmmmm
So, 32 of 100 is simply 1.6 stars.
Formula 32 / 20 = Number of stars.
Wow. I don't see why this article is on slashdot, what am I missing?
Modesty is one of life's greatest attributes
fucking confirpmed:
Slashdot can now link to past duplications. It's really quite funny because I scrolled down and they linked directly back to the same story. In related news the same story we will now provide the same story slightly edited..
Ooo man the floppy drive is broken. No wait. The computer is just upside down.
[Fuck Beta]
o0t!
are abou7 7000/5
Interesting to see that what would just involve reading a few lines of code on open source software takes a whole complicated statistical analysis on closed source software...
Any idea why they would pick 100? There doesn't seem to be any obvious reason to do this over 0 to 5. Memory considerations would peg it to 2^n-1 values wouldn't they?
m0nstr42.blogspot.com
iTunes? Is anyone still using that old thing? I now exclusively use Songbird, even though it's still alpha.
Tsk, and I thought this was Slashdot... bleeding edge, people!
They could have just been leaving their options open for future expansion of the software. Indeed, it would be better for compatibility to already use the 0 to 100 range.
Suppose in the future they added support for giving a song a percentage ranking. It's quite easy to do with the system they're using now. Data from the new software could be used with the old, as the old software is already expecting and can cope with a value in the 0 to 100 range, even if it does so slightly differently than the future versions.
Cyric Zndovzny at your service.
Speaking of crashing music programs, my friend managed to crash an old DOS metronome program by setting the time signature to 4/0 time (normal time signatures look like 4/4 or 3/4 or 7/12 etc). It divided by zero and crashed. I found it pretty funny at the time.
By the looks of things they need to use another program to do linear regression on data. That line of best fit would make my old physics prof cringe.
It's quite obvious that it doesn't impact on a normal user's experience. I thought that went without saying, but apparently you completely missed that point.
So in the end it doesn't crash the program, nor does it lead to incorrect output or corrupted data, or anything of that sort. Thus it is not a bug.
Cyric Zndovzny at your service.
Why is the rating system linear instead of exponential. I would like a song that has one more star than another to play TWICE more often.
On the other hand, even I think this is too boring to tell other people.
"The White House is not an intelligence-gathering agency," -- Scott McClellan, Whitehouse spokesman.
Can somone kindly tell me 'why'?
This earth shattering revelation will truly shake religions and politics alike throughout the world.
I'm stunned.
Important stories like these is what make Slashdot so great.
Beware: In C++, your friends can see your privates!
Maybe not on topic, but in Norway iTunes is getting investigated by the goverment for violating various consumer and marketing laws. Will be interesting to see what the outcome will be.
Why did Apple choose multiples of 20? Isn't 0, 1, 2, 3, 4, 5 a lot easier to work with? It certainly makes the most sense.
here is some more advanced math for you: iTunes = Teh Suck.
In the current Windows version of iTunes (6.0.2.23), in Preferences there is a tab for Playback. And in there is a section for Smart Shuffle.
/didnt rtfa
The first control is a 3-setting slider for "more likely, random, less likely". The description below says "Smart shuffle allows you to control how likely you are to hear multiple songs in a row by the same artist or from the same album."
Below that are three radio boxes, labled Shuffle: ( ) Songs ( ) Albums ( ) Groupings.
Is it possible that this 100 scale rating system might be more fine grained to take these Smart Shuffle user settings into account? Perhaps that might be why it isn't simply the integers 0-5.
Somebody has too much time on their hands...eh?
But 'science'??
my password really is 'stinkypants'
introductory algebra is new to you. Sorry, it's not people who think this is boring that should be looking at Disney.com. It's people who think this is interesting that should be posting somewhere that's the Disney equivalent of Slashdot.
I don't use itunes, and while I realize many people do, this seems pretty mundane. It might be interesting to those who care - obviously there's some of you out there - but this hardly seems worthy of Slashdot. I'd expect something more hackerish along the lines of "NSA backdoor found in Windows". Math wise I'd expect something about a newer, better way to find prime numbers, or a new encryption algorithm. Not how cool it is that instead of 0-5, they use 0-100!!! And you can specify half a star. BAD ASS!
Just my useless, anonymous two cents.
You can do half star ratings in iTunes for Windows as well. Works exactly the same way.
g rammery/
Scripts to do it can be found here: http://ottodestruct.com/blog/category/geekery/pro
- Give a man a fire and he's warm for a day, but set him on fire and he's warm for the rest of his life.
"that not only do non-whole, non-half values (e.g., 47) not increase the amount a song will play, but neither do the half-star ratings (with the exception of .5 star)."
I know this article is about DRM-filled software, but do we really have to be this negative towards it?
Move Sig.
How in the heck does a stupid story like this make it onto slashdot, let alone the science section, what crap!
47 really really IS the WEIRDEST number... Really. There's even a group dedicated to it:
The 47 Society
So TFA comes as no surprise...
RS
Shoes for Industry. Shoes for the Dead.
If your song gets a rating of 42, you can expect to hear only that song while you ponder the meaning of life.
Proof by very large bribes. QED.
{
FTFA:
"I remember when I first discovered MP3 files back in the early 1990s..."
I call BS. From wikipedia:
--"July 7, 1994 the Fraunhofer Society released the first software MP3 encoder called l3enc."
--"In the first half of 1995 through the late 1990s, MP3 files began flourishing on the Internet.
so basically before mid-1994 there were no mp3s. There were other things (like mp2?) but mp3s really didnt exist since, duh, u kinda needed a encoder in order for mp3s to be created.
ok so this is minor and really has nothing to do with the article but i just hate it when you show someone some cool neat thing you found and they're all like "ya i discovered that back in forever ago" especially when it didnt even exist yet .
} //END RANT
mod me up or down i dont care i bleed karma
my karma will be here long after I'm gone
so basically before mid-1994 there were no mp3s. There were other things (like mp2?) but mp3s really didnt exist since, duh, u kinda needed a encoder in order for mp3s to be created.
Dude, he said "...MP3 files back in the early 1990s...". In case you didn't notice, 1994 is part of the early 90's, in that it's before 1995.
Not sure why you're bitching about this, it doesn't seem to be contradictory at all . . .
Can I get an eye poke?
Dog House Forum
If you read my article on the Rating Systems you'll see that 5 point rating systems are a poor way to do ratings. In our analysis, only 10% of raters use the bottom two star in a 5-point scale, and only 2% use the lowest rating of 1 star. The median of the 5-point scale is actually the fourth star, with a neat bell curve arranged around it. In my own personal use of iTunes, I've forced myself to be much more consistent and lower with my ratings. 1 star means that it has been rated, but in general I don't like it and don't play it. 2 stars means average, which means play it some of the time. 3 stars is above average, and play it more often, but I can get bored with these. 4 is good, and can basically listen to these regularly without getting tired. 5 stars are exemplars -- I try to find more songs like these.
Didn't you hear? You can't claim wikipedia as a source!
...jk, i <3 wikipedia.
"Banking establishments are more dangerous than standing armies." -Thomas Jefferson
IMO 1994 counts as the mid 90s. I would hardly think discovering MP3s less than 6 months before the end of the first half of the 90s counts as early 90s.
He's bitching because he had to reach into wikipedia to back up his assertion, instead of citing anything from personal experience. In other words, he was likely listening to Barney sing back in the early 90's.
1) iTunes doesn't play half my collection, because there's no ogg support. Winamp plays ogg, but it's got nothing for library management. Somehow, Songbird managed to pull off ogg support before they were even released, and its shooting for the basic skeleton of features iTunes provides as well.
2) Songbird helps you buy music from a variety of stores, rather than just one. I think eMusic is better than the iTunes Store, but even if it isn't, I don't want my car telling me where to buy gas; I don't want my music player telling me where to buy music.
3) If you're going on the web to look at media, going on the web with a media player makes some sense. Reading music blogs with Songbird feels a lot more sensible, you can stream recommended mp3s through a different pane, without leaving the page. It feels a lot more natural. Granted, this is one miniscule fraction of the things we do online, it probably wouldn't make a lot of sense to use Songbird to check your Gmail account... except whatever advantages are achieved from not having to open a separate program if you're listening to music and just want to jump online to check something really fast.
4) It's still just a proof of concept. It's not finished, and there are no extensions for it yet. So the fact it doesn't stack up to iTunes yet is a non-starter. You want less bugs? Wait until it's even in beta. You want ratings? Maybe they'll be incorporated, maybe they'll be an extension. You want a smarter shuffle, one that shuffles by artist rather than by track, or one that "sticks" to a genre or tempo? Extensions. You want it to automatically download the mp3s off your favorite music blogs every day? Extensions. You want anything new in iTunes? Go whine about it to your preacher, because I've sent in my complaints to Apple, and Apple's not listening.
I hate it when someone finds something cool neat thing and then someone else attacks them on something totally irrelevant. You feel big now? Hope it helped your self esteem to reveal the dark fact that he in fact DIDN'T listen to mp3's in the early 1990's. I think you will successfully convince the jury with that one. Lets break out the torches and pitchforks and lay siege to his house. Or we could just be impressed with a pretty cool analysis of a software program and over look his poor estimation of when he started listening to MP3s.
I have to say that while the aim of the article is quite admirable, the author does a horrible job of analyzing the data he collects, and he completely misses an anomaly of interest.
His conclusion that "only for tracks rated 1-19" do ratings of finer granularity matter is bogus. Ignore figure 5 and look at every other figure in the article. It's not linear from 1-19 either. It shows the same step behavior as the rest of the graph. He mistook an anomalous edge case for some sort of liner relationship. The *actual* interesting piece of information that can be taken from the data collected is that, whether intentionally or not, the algorithm Apple uses has an odd (again, I don't say erroneous since it's possible that it's intentional) edge case. The play count difference between two songs where one has a rating one star higher is around 2000-2500. However, between no stars and one star, the rating difference is 3700. Without more information (and a Mac of my own to investigate myself), it's not clear what's going on. The author states that the 0-100 rating gets integer-divided by 20, which makes sense except that that's not what the data in the graph reflects since a rating of 0 yields a significantly different playcount from a rating of 1. In fact, neither a strict ceiling or floor explains the data generated, especially since the rating is discretized to a five-star rating.
Here's my point in summary. If iTunes integrally divides or floors the rating, then where does the extra step when the rating is between 1 and 19 incluseive come from?
Itunes. But I'll tell you what I don't hate, The spicy italian sandwich from Subway. 2 layers of salami and Pepperoni topped off with 4 slices of american cheese. Put a little mayo on there, some lettuce, tomatoes, and black olives, and you've almost got yourself a sandwich. Next, have them cover the sandwich in salt and pepper, then spray some oil and vinegar over the top, wrap it up and there's dinner. I think we can all give thanks for a sandwich that good. mmmmm mmmmm