This file lists long-time goals that are not fixable quickly. Issues that are easily
fixable should not be recorded here, but should rather be fixed in the code.

Structural:

- LibDA as a separate project
	* Also playback and Vorbis + MP3 decoding
	* Mixer and samplerate conversion inside
- Get rid of ScreenManager in favor of recursive approach?
- Switch to Clanlib or SFML?
- Switch to GStreamer? (both the ripper and the game)


Big features:

- Add compose screen (using features discussed on mailing list)
- Multiplayer duet (needs XML or MIDI song format)
- Recording performances
    * Internet highscore
- Mic effects (reduce volume for bad singers, perfect pitch, reverb, etc)
- Ingame Singstar DVD support (integrated ripping + ffmpeg feeding + XML song format)
- Dance game [WIP]


Features:

- Latency calibration
   * Audio roundtrip by analyzing the sound captured by the microphone
   * Controller/audio latency by playing blindly (listening to the beat only)
   * Audio/video by playing visually without audio? (needs controller/audio calibrated)
- Better audio card configuration ingame (needs PortAudio backend) [WIP]
- Allow playing a single song given as commandline argument (no menus)
- Kiosk/arcade mode (disable exiting game and limit other functions)


Dance features:
- Colored arrows for timing
- BGCHANGE support (background videos)


Text rendering bugs:

- Correct vertical alignment
- Use dimensions to specify bounding box (possibly init from SVG)
- Lyric hilighted should be rendered in front of the others


New config menu:

- Multiple items visible at once
- Modifying strings and stringlist items.
- Predefined options from config schema and libda device enumeration


SS Ripper:

- Compress into Vorbis Ogg and H.264 MKV


Bugs:

- On Compiz switching from fullscreen to window makes the window maximized even though the viewport stays small

- Segfault when browsing songs in song browser while loading (segfaults before loading finishes), quite rare

- Bitmap surfaces are often loaded only partially, loading stops at a middle of a scanline.
   * Affects at least PNG format, maybe not others
   * The rest appears black or transparent (at least most of the time)
   * Reloading often causes different results on the same image
   * Different images may also stop at the exact same pixel
   => out of memory condition or something?

- Instrument data files + catch

