Re: [Rhythmbox-devel] Improved random patch (bug 163196 ish)

["Dr. David Alan Gilbert" <gnome treblig org>]

* eric casteleijn (thisfred gmail com) wrote:
> >that's funny. I've investigated the same thing, though without getting so 
> >far as to submit a patch.
> >Especially I was interested in lowering the probability of very recent 
> >songs being replayed.
> 
> Slightly OT/tangential: In my (non-random) plugin 'autoqueue', I have a 
> configurable number of days that a song is 'blocked' for, and the same 
> for artists. That last part ensures even more variation.
> 
> Maybe allowing the number of days to be user configurable is a good 
> idea? Two days seems rather arbitrary, and might be too long for some 
> and too short for others. Maybe a logarithmically declining (meaning 
> very sharply at first then levelling out) scale would be even better, or 
> is that what you're doing? Hmm, I may want to do that...

Indeed, the thing is the time for 'very recent' would vary depending on
the size of your library, your usage patterns and your memory!  In my
patch there is a slider so you set it in days.  The penalty is a simple
divide of the weight (by another slider set amount) for those that
are in the recent period.
I *think* the power/log arrangement I use for the normal non-recent
period also leeds to a rapid tail off from recent times.

Dave
-- 
 -----Open up your eyes, open up your mind, open up your code -------   
/ Dr. David Alan Gilbert    | Running GNU/Linux on Alpha,68K| Happy  \ 
\ gro.gilbert @ treblig.org | MIPS,x86,ARM,SPARC,PPC & HPPA | In Hex /
 \ _________________________|_____ http://www.treblig.org   |_______/