extragear/multimedia/amarok/src/playlist

[Leo Franchi]

On 30 Oct 2008, at 12:34, Nikolaj Hald Nielsen wrote:

>> I still insist that since grouping is a non-essential function of  
>> the playlist
>> view, then the code should *always* err on the side of not grouping  
>> albums
>> that should be grouped (false negatives), rather than grouping  
>> albums that
>> shouldn't be grouped (false positives).  Does the current code  
>> group together
>> albums that shouldn't be grouped?  Are these cases fixable without  
>> resorting
>> to more false positives?
>
> No, the current code ( still using pointer comparisons ) does not
> group together albums that should not be grouped, but the amount of
> things that should be, but are not, is considerable. Right now I am
> merely going on what _feels_ the least broken, and to me that is
> clearly using the album names. Even if tracks from similarly named (
> but distinct ) albums are incorrectly grouped together ( something
> that can be reduced significantly by checking the album name if
> available ), it will still be immediately clear to the user _why_ this
> happens ( hey, they have the same album name.. that is a bit silly but
> sort of make sense. i.e. minor bug ) , as opposed to now where the
> playlist feels broken as it does not group a bunch of cases that, from
> a user perspective, should obviously be grouped ( Oh, adding this book
> from librivox does not group the tracks together, even though they all
> have the same album name. Hmm, this album I just added from the file
> browser is not grouped either... That looks like a big bug in the
> playlist  ).
>
> As I have said many times, I am more than willing to do a popper fix,
> post 2.0.0.

I would say grouping is one of the essential features of the new  
playlist. The way it has been graphically designed relies on coherent  
grouping for it to actually look good.

Second of all, i think our priority is to give the best experience to  
our users. As all have agreed, this is NOT a long term fix. this is a  
temporary fix until we can properly fix it post 2.0. What is most  
important is to release a version of Amarok that works the best for  
*almost all* cases, instead of one that is technically correct yet  
breaks for the user in a bunch of ways.

cheers,
leo

---
Leo Franchi				(650) 704 3680
Tufts University 2010

lfranchi at kde.org
leonardo.franchi at tufts.edu