Abstract idea of Song instead of pure file management

[Casey Link]

Hi,
First, your idea is a good one, and while not the first time we've
heard it (it's popped up  several times over the years), it's
definitely the most thoughtful proposal I've seen for this feature.

In Amarok 1.4 this feature simply wasn't possible because the
underlying architecture wasn't readily adaptable to this second layer
of abstraction. In Amarok 2 this architecture has been completely
rewritten, and with some work this feature is definitely possible.

On Tue, Dec 2, 2008 at 11:13 AM, Jud Craft <craftjml at gmail.com> wrote:
> I think that's right.  I believe you're talking about a type of
> song-identification based on algorithmic analysis of the song itself?

Yes, that's what Jeff was referring to.

> I don't mean to suggest that's not useful, but my interest was in
> grouping together files that map to the same song, as opposed to
> dynamically comparing two files to see if they "might be" the same
> song.

Manually grouping identical-different-format songs together is
important, and will likely make an appearance in this feature. This
presents a large user interface issue. A user shouldn't have to
manually group his songs, just like he shouldn't have to manually
group his albums under their respective artists. In the case of
grouping albums under their artists, Amarok reads the tags of the
track's and displays them accordingly. The only way to achieve
automation of your type of song abstration is with some sort of song
identification.

Song identification can also be used for many other use cases, but
that's besides the point.

So, while automatic identification isn't required for song
abstraction, it will definitely be used in Amarok.

I think this is a very important and revolutionary feature;  hopefully
it will make an appearance in the 2 series. If I find the time among
my other tasks I plan to take a stab at it.

Casey