A better default scoring script for Amarok 2

[Andreas Mützel]

Hi,

2008-12-14 Karthik Periagaram:
> With these changes, the score algorithm change is complete. The new score 
will 
> agree with your rating more and will approach 100 as the song playcount 
tends 
> to infinity. Thus, it is a better algorithm to calculate how much the song 
is 
> liked by the user. 

in general I like the idea of not "punishing" a track too much if it is 
skipped a few times. But from my point of view there would be a serious 
problem if I understand the "will approach 100 [or at least 90] as the song 
playcount tends to infinity"-part correctly:
At least on amarok 1.4 i tend to have about 1 week of music in my playlist in 
random mode. Usually i simply skip songs i don't like and I use the scores 
mainly for putting music on my mp3 player because they tend to be high for 
songs I don't skip and low for songs I skip often.
So by sorting the playlist by score and transferring like the 100 highest 
ranked songs to my player I get a pretty good choice of music when I'm not at 
my PC.

Maybe there are more reasons for wanting to semi-actively "punishing" a track 
when in random playlist mode, for example when using the scores in a biased 
playlist. (-> "I don't like that song, so if I skip it a few times it will be 
excluded from the playlist in the future.")

I know, sometimes it is not desired to score a track too badly because it is 
skipped once, but with the formula you proposed it would still be possible to 
reach a score of 90 for a song that matches my dynamic bias but is always 
skipped...

Somehow I like your idea, but I hope I could show the problems I see in a 
useful way...

Regards,
Andreas