Version: 2.0 (using KDE 4.1.3) Compiler: gcc (GCC) 4.1.2 (Gentoo 4.1.2 p1.1) CHOST="x86_64-pc-linux-gnu" CFLAGS="-march=nocona -O2 -pipe" Installed from: Gentoo Packages problem is that I can't really tell how amarok calculates my playlist making it near impossible to tweak. I really don't know what it's doing. what does it do when I have proportional bias 75% genre rock and 60% genre pop or even more confusing what if I have a 80% fuzzy bias on year 2008 and a proportional and an 80% proportional bias on classic rock (assuming all tags are correct). in the spirit of the first example lets say 60/60 does that end up being 60/40 using the first value it comes across? 40/60 or 50/50? one suggestion was offer a pie chart... some kind of chart might be nice. another would be that in the case of proportional don't allow the sliders to add up to more than 100% I'm not sure how that affects fuzzy though.
You might want to check this out: http://kopophex.blogspot.com/2008/08/introducing-fuzzy-biases.html
IMHO it is fine the way it is. And definitely no fancy diagrams! It is also a bad idea since you might want to have two completely unrelated biases like "from 1970" and "rock". They don't need to match up to 100%. And really: Needing this to match up to 100% screams OCD. WONTFIX from me. Sorry.
doesn't really matter if it adds up to 100% it's just an example, it should if it really is percentage based but maybe it's not. it could be point based. which would mean 60/60 adds up to 120 and each would have 50%. I'm just suggesting that the UI somehow needs to better reflect what's actually going on. which is why I've tried to provoke a discussion because I don't have a great thought on how to fix it, only that I think it could use improvement. (also these features seem to be very similar to something I suggested when amarok 2 was in pre-alpha, for all I know it's from my idea).
Have to agree with Caleb here. The "concept" is pretty and powerful and amazing but as an end user it's absolutely confusing! How do I select all songs from the 1990's that are *either* R&B or Soul but *not* from 1995? It was very easy to do before: AND conditions: year between 1989 AND 2000 year not equal to 1995 OR conditions: genre = R&B genre = Soul
*** Bug 199479 has been marked as a duplicate of this bug. ***
In my opinion, any interface that allows: A: 99% B: 10% is simply borked. I don't care if it behaves the developers want it to... there's simply no way for the user to know what that behaviour is. I have actually spent a lot of time trying to figure out what it really does (just by looking at the playlists produced) and it still looks like black magic to me. For example, it turns out that: A: 99% B: 1% generates playlists in which 1% of the songs don't match either A or B. That looked like a bug to me, and I filed it as such (bug 198836), but apparently it's <i>not</i> a bug. So I am completely mystified by this UI. It would be marginally better than the current situation if the behaviour were documented *in detail* in some kind of help item accessible from the UI. At least then a user would know what's supposed to happen and could file a bug if the observed behaviour were different. But as it is, the user is left to navigate a maze of twisty passages, all alike. Or something.
we are very very aware of the issues with the dynamic playlist gui. we understand how the GUI does not at all reflect how the bias solver actually works, and is really confusing. yesterday, here at akademy we had a group discussion and worked out a plan to go forward, so a much better UI + backend should at some point be coming.
I think we should retag to 2.3 because we are in feature freez
What about using 5 steps instead of 100 single steps? Currently: 1..100% Proposed: "None" "Few" "Normal" "Many" "All" (or similar terms). So it wouldn't cause confusion because the math of adding up values isn't right.
Without taking stance what is the best scheme, in the current situation I'd like the first bias to be 100% instead of 0% when it is added. I mostly use just one bias and it takes more work on the mouse to pull the slider up.
There seems to have been some change in the way the dynamic playlist slider values are interpreted between the releases for jaunty (April) and karmic (October). Settings that used to give me what I wanted now do something different. In jaunty (don't recall the Amarok version; 2.1, I think): Rating == 5 [100%] Rating == 0 [1%] gave me an occasional unplayed song (I can't say that I understand why, but I was told to do that, and it worked). In karmic (Amarok 2.2.0) the same settings never give me any unrated songs. Isn't the algorithm documented clearly somewhere so I can at least understand how it's supposed to work? I find it extremely frustrating trying to guess what was going in the developers' heads when they came up with this interface -- especially since the behaviour seems to have changed at some point in the last six months.
Rick, do you have input on this?
2.3.1 is in feature freeze -> 2.3.2
I am just working on the new dynamic playlists. It should be easier to understand and also prevent duplicates. If you want to have a look see the "dynamicplaylist" branch on git.
New dynamic bias has a "partition" bias. I think it's much easier to understand, however the "fuzzy bias" was discontinued in the rewrite.