merge request: Automated Playlist Generator

[Colin Kern]

Disclaimer: I am not an Amarok developer, just a user.

I would like to have this feature, but as you say the UI has a lot of
room for improvement.  The biggest thing that jumped out at me was
naming all the rules.  Why is that necessary?  Just displaying the
value and constraint is descriptive enough, I think.  Another issue I
have (and this is also an issue with dynamic playlists) is the
strictness value is not very clear.  If it's doing the same thing as
the fuzzy bias in dynamic playlists (controlling the standard
deviation) then referring to it as a percentage from 0 to 100 doesn't
really make much sense and it isn't clear to the user what it's doing.
 I'm not sure how hard this would be, but I think a good way to
present this would be to show a margin of error, so for example if
they were setting the duration of the track, the text of the slider
might say "+/- 0:30".  Of course, since the math involved here is a
normal distribution, you can't say that definitely, but maybe have
that number be the span of 2 standard deviations (95% confidence).

If I have some spare time, I will clone your repository and play
around with the feature some to see if I get any more ideas.

Colin


On Tue, Mar 16, 2010 at 10:25 PM, Soren Harward <stharward at gmail.com> wrote:
> I've created a merge request for the Automated Playlist Generator:
>
>     http://gitorious.org/amarok/amarok/merge_requests/148
>
> If you want to play with the code, you can get it from the "generator"
> branch of my git repository:
>
>     http://gitorious.org/~stharward/amarok/stharward-work
>
> I've also uploaded a demo video to the AmarokStudios YouTube channel
> (thanks, Lydia, for helping with this):
>
>     http://www.youtube.com/watch?v=h6VEsvQNnjQ
>
> I wrote the APG because the current dynamic playlist/bias system still
> hasn't achieved feature parity with the Smart Playlists of v1.4, and
> the algorithmic parts of the bias system have some, IMO, serious
> design flaws.  I will admit that the UI of the APG (particularly the
> preset editor) is not as straightforward to use as the UI for the bias
> system for dynamic playlists, and expect to make some changes based on
> user feedback.  However, I think the current Preset Editor is no more
> difficult to use than Smart Playlists in v1.4 were, and the cost of
> getting over the learning curve is more than offset by the fact that
> the APG actually does what you think it should based on what you see
> in the UI, unlike the current bias system.  In other words, a
> moderately complex UI with a stable backend that performs correctly is
> preferable to a simple UI with an unstable backend that performs
> unpredictably.
>
> --
> Soren Harward
> _______________________________________________
> Amarok mailing list
> Amarok at kde.org
> https://mail.kde.org/mailman/listinfo/amarok
>