[Rhythmbox-devel] A thought for the future...

[Douglas McMorris <virage83 mail utexas edu>]

This is by no means something to be considered for the next stable
release, but i think it would be nice to improve integration with
nautilus and the desktop in general.

If we want Rhythmbox to be THE gnome music player, then it would be nice
to have a way of playing audio files without adding them to the library.

Heres my idea for how this could be done.  First of all, I think that
having a small view, similar to the one itunes, wmp, and others have
would be nice.  It would also work well for album covers and possibly
goom (see attachment).  the way i invision this working is that if you
launch a file from nautilus it opens up the small view and plays the
file... what the user expects to happen (imo if its registered for audio
files).  The file starts playing and they are asked if they would like
to add this file to the library (doesn't ask if its already added) and
that dialog would have a checkbox to always add and not ask again, or to
never add and not ask again (something like "remember my decision"). 
There could also be a menu item or a toolbar button that would appear in
the situation.

In the long run i think this would allow for rhythmbox to be more
integrated with with the desktop.  most users would be asking them
selves... i just want to listen to the audio clip i downloaded... why is
the music library coming up and not even playing it???  which is where
we're at now.  On that note, this would also provide a nice light
interface for playing iRadio from browser links and such.

Sorry for the long email, but I've been thinking about this stuff for a
bit and have a lot to say I guess... or ramble about... you decide
that..

two more things...

It occurs to me that most of the code complexity in rb is the library
its self which (with the way we use it) is basically a dumbed down
database.  it seems like using a more mature library for doing a
databased approach would do a lot for us:

1.  dump a lot of the complexity into prove code that we aren't having
to maintain

2.  speed up the speed of queries

3.  make it possible to have almost any imaginable for of query. 
multiply genres, albums, or artist selected, search for a title with in
a certain genre.  things we already want to do with smartplaylists and
would leave the door open for tons of ideas later.


one last thing.


Something badly needs to be done to the documentation work on the code. 
i tried learning the code base a while back, but quickly got discouraged
b/c i had no idea what sections were doing and there was no
documentation on what was going on either.

I've been following rb for a while and it seems like there have been 2
spurts of interest, hacking a new UI... then a lack of any activity for
a while.  look at the changelog no commits in almost a month.  it seems
to me that most of this is b/c of the lack of good code doc work. 
whatever we do from the point on needs to be well documented or this
project will live only as long as those who wrote the code are
interested.

please understand that none of this is intended as a flame.  i am
personally looking forward to spending some time hacking on rb my self,
currently i don't understand the codebase so i'm pretty much useless
right now.

let me know what you think... i mostly want to stir discussion and get
ideas out.

i think the biggest thing rb is lacking as this point is a concise
direction and thats something that needs to be discussed and nailed down
(IMHO)
 
-- 
Douglas McMorris <virage83@mail.utexas.edu>

rb-small.png