* Problems with GNOMAD:

  Make it possible to edit UpPeRcAsE in metadata beacuse it
  isn't right now goddamit.

  Codec type "mp3" instead of "MP3" on transferred files!
  Notably RMB Mitternacht. (Haven't happened again...)

* Libnjb:

  No problems except for stability

* Problems with the USB protocol spec:

  The commands 0x29 and 0x2a are not documented.

  Request: C3
  Command: 29
  Name: Get time
  Description: Gets the time of the internal clock of
               the jukebox.

  Request: 43
  Command: 2a
  Name: Set time
  Description: Sets the time of the internal clock
               of the jukebox.

  Their datatypes can be documentet with a brief look at
  time.c

  (p 44)

  The data file tag size sent in the data phase includes
  the four padding zero bytes.

  --- Already sent to John: ---

  Status codes (p 15):

  Add for playback:
    0x40  NJB_ERR_QUEUE_ALREADY_EMPTY

  EAX block structure (p 13):

  The blocks named:
    number of spatializations, nspat
    active spatialization
    nspat x spatialization frames
  Is actually:
    number of headphone modes, nphone
    active headphone mode
    nphone x headphone frames

  Also it should be noted that effect, headphone,
  and rear frames are actually just a sequential 
  list of strings with the names of the mode at
  indexed position.

  Adjust sound (p 53):

  The table of effect bytes:
    0x01 -- Volume value is between 0 and 100
    0x08 -- Amount of current EAX effect a value between 0 
            and 100
            note that a change actually changes the
            amount of ALL effects, not just the current
            one. The EAX structure will hold the same
            value for all effects afterwards. (Atleast
            with Firmware 2.93.)
    0x0b, 0x0c -- these must have a meaning too, but I
            just cannot decipher it... Tried all kind of
            input I could think of.

* Problems with LIBNJB:

  Something about the memory allocation for playlist iterators
  and playlist tracks is still causing me serious pain now and
  then.

  Something weird with delete track from playlist. (?)

  These functions seem not to like if you (softwarewise) captures/opens and
  closes/releases the jukebox too much:

  * NJB_Delete_Track
  * NJB_Update_Playlist
  * NJB_Delete_Playlist

* Things that would be good in LIBNJB:

  If
  playlist_deltrack (playlist_t *pl, unsigned int pos);

  Took a track ID instead if a position, would be good:
  playlist_deltrack (playlist_t *pl, u_int32_t trackid);

* New stuff:

  * Keybindings to all menus (including popups)
  * EAX settings?
  * Unicode support (Good for NJB3) <-> ID3v2 (Argh!)
  * Add support for YEAR tag in the xfer code. (?)

  DAVE:

  * migrate to a better config situation
  * queue showing -- even w/o full queue
  * queueable
