Re: [Rhythmbox-devel] Re; Ratings feedback



I suggest ensuring that there are plenty of options in some sort of 
dialog box that allow the user to choose what RB does in response to 
certain actions.

Examples (probably would need to come in a more intuitive format, but 
here's some ideas):

Lower a song's rating when skipped within this timeframe: [5 seconds/10 
seconds/30 seconds/60 seconds/Never (blanks out the following option)]
Decrease a song's rating by this much when skipped: [2 stars/1 star/1 
star if skipped twice/No change]

Random play:
Play 5-star songs [100%/80%/60%/40%/20%/0%] of the time.
(and so on for each star rating)

And so on, for everything useful that we can think of.

Everyone's going to have a different opinion on how they want their 
ratings treated, let's give them the ability to do it how they want.

Jeff Perrin wrote:

>in7y118@public.uni-hamburg.de wrote:
>
>>Snip
>>
>>What this behaviour should mean to Rb:
>>- Listening a song to the end means I liked listening to it. I'd probably like 
>>listening to it again.
>>- Skipping it means I don't want to hear it right now. Maybe later, but most 
>>likely not "soon".
>>- Selecting a song explicitly means I like it a _lot_. I probably wouldn't 
>>mind if it was played three times in a row.
>>- Adding a song to the library means I like it. I wouldn't add it if I didn't 
>>want it to be played.
>>
>>What I want Rb to do:
>>- Make it so I never need to use F12.
>>
>>Benjamin
>>  
>>
>
>I'm responding to Benjamin, but please keep in mind that I've never tried the .7 
>versions of Rhythmbox, so perhaps I'm just asking for something that already 
>exits...
>
>Basically, I think that the adaptive ratings are a pretty cool thing to have, 
>however I'd probably never rely on it myself.  What Ben has just described seems 
>to me to be an around-about way of self-rating your own music collection.  Only 
>instead of explicitly saying "this song should have a rating of 3 stars", I have 
>to play the song all the way through.  Instead of saying "this song is like, 
>totally rad, it should be 5 stars", I have to explicitly search for the song.  
>Instead of saying "this song sucks ass, it should be one star", I have to press 
>the 'next' button before the song completes playing.
>
>When I first installed iTunes on my Windows partition, I would explicitly rate 
>each song as I imported it into my collection.  By the time I had my entire 
>collection ripped onto my hard drive, I had about 500 or 600 songs that I 
>*really* liked and that I'd like to listen to no matter what.  There is no way 
>that I could have achieved this listing of bitchin' 4 or 5 star tunes in the 
>amount of time that I did without explicitly rating the songs myself.
>
>The *one* thing that I felt was missing when I loaded my iTunes library into 
>Rhythmbox was the auto-playlist allowing me to only listen to songs that I had 
>rated at 4 or 5 stars.  This is a great feature.  I personally don't care how 
>these ratings are generated, however I fail to see how the auto-ratings can 
>possibly be faster and more accurate than me personally going through my songs 
>and explicitly choosing which ones are rad and which ones suck.  The auto-rating 
>feature seems like a great academic exercise, however it's something that I'd 
>personally probably never rely on.  I already *know* which songs suck.  I 
>shouldn't have to listen to part of them before Rhythmbox knows this.  As a 
>user, I'm much more happy to have the ability to rate my own songs.
>
>What I guess I'm looking for is a reason why adaptive ratings are better than me 
>explicitly telling Rhythmbox what I do and do not like.  To me, the adaptive 
>ratings feature looks like it's an excuse for somebody to write some cool code, 
>instead of something that most users will actually need and understand.  I don't 
>really care if the feature exists, as long as I can tell Rhythmbox what I like 
>myself, and then generate an auto-playlist based upon my own ratings.  Please 
>excuse me if this is something that I can already do... I'm just worried that 
>the adaptive ratings will take the place of my own explicit rating of songs.
>
>>Quoting "jorge o. castro" <jorge@whiprush.org>:
>>
>>  
>>
>>>Derrick Ashby wrote:
>>>    
>>>
>>>>I think the point is that this should be a thing that the user decides for
>>>>      
>>>>
>>>>themselves.  You prefer 3 because you only rip songs you really like.  I
>>>>      
>>>>
>>>rip 
>>>    
>>>
>>>>every CD I buy, and I download tracks also: For me 0 is not rated, 1 means
>>>>      
>>>>
>>>I 
>>>    
>>>
>>>>don't like it, 2 means it's OK, 3 means Good, 4 V. Good, 5 Excellent. 
>>>>Starting at 3 is a waste of 2 stars.  I'd like to be able to turn Adaptive
>>>>      
>>>>
>>>>Rating off altogether.
>>>>
>>>>      
>>>>
>>>I've got some feedback on adaptive ratings as well:
>>>
>>>I also rate my music in the same way Derrick does. I've been "training" 
>>>  my rb since .7 came out and have made some observations:
>>>
>>>The IMMS-like ratings doesn't really apply to the 5-star model of 
>>>rhythmbox very well. From scanning my xml file a good number of songs 
>>>are "2.8" or somesuch. This means for it to be totally accurate partial 
>>>stars would need to be implemented, but I'm not sure if that's a good 
>>>idea or not. Obviously, the larger a user's collection is, the longer it 
>>>will take for the "training" to become effective. During that time the 
>>>user might just say "ugh, I've given the thing 3 weeks and it still 
>>>thinks all my songs are three stars, this is worthless."
>>>
>>>I've been doing some thinking about my music collection and listening 
>>>habits, and there are certainly songs I keep around for "completeness", 
>>>ie. I like an album and want to keep it in my library for the sake of 
>>>completeness, but at the same time it has certain tracks that I don't 
>>>care for.
>>>
>>>We could probably argue that Derrick's rating is what most users think, 
>>>but someone might say "well, if I don't like a song I just don't keep it 
>>>at all." Perhaps a more balanced default behaviour is needed, after some 
>>>thinking I propose the following:
>>>
>>>1) Adaptive listening should be off by default.
>>>2) A user's manual rating should override the automatic rating.
>>>3) All imported songs should start off as 0. If a user skips the song or 
>>>it is shown as unfavorable, then it remains at 0.
>>>4) Songs that are shown favorable should start rating up according to 
>>>listening habit.
>>>
>>>I've come to these conclusions based on feedback from both my experience 
>>>and some people that I know. Basically, when I hit "random" on my music 
>>>player, I want rhythmbox to choose random songs for me. While I chose 
>>>"random", I would argue that most users really mean "pick a good 
>>>selection from favorite songs, but mix it up enough to surprise me on 
>>>occassion." If you think about it, a "true random" feature isn't much use.
>>>
>>>On top of that, I don't know a single person that takes the time to sit 
>>>there and spend hours rating all their songs. If anything we tend to 
>>>take the lazy way out and say "Well, I'll just sort by play count to 
>>>figure out what I want to listen to."
>>>
>>> From a user perspective, we should be letting Rhythmbox do the hard 
>>>work and figure out what my favorite songs are, I mean, it's the one 
>>>sitting there doing all the playing, it should keep track of the stuff 
>>>and figure out a decent random list that will keep me interested.
>>>
>>>These are just some observations I've had, I have no idea what the 
>>>"optimal solution" is, which is why I propose it be off by default, but 
>>>at the same time, from my long experience with XMMS and IMMS, given 
>>>adequate training the stuff really does give me a random playlist that 
>>>totally rocks.
>>>
>>>-jorge
>>>_______________________________________________
>>>rhythmbox-devel mailing list
>>>rhythmbox-devel@gnome.org
>>>http://mail.gnome.org/mailman/listinfo/rhythmbox-devel
>>>
>>>    
>>>
>>
>>
>>_______________________________________________
>>rhythmbox-devel mailing list
>>rhythmbox-devel@gnome.org
>>http://mail.gnome.org/mailman/listinfo/rhythmbox-devel
>>
>>
>>  
>>
>
>_______________________________________________ rhythmbox-devel mailing list 
>rhythmbox-devel@gnome.org http://mail.gnome.org/mailman/listinfo/rhythmbox-devel
>  
>




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