TODO:
----
-* Get the echo instances to behave consistently
+* Get the multiple choices in both local and remote CDDB to use the same
+ functions.
+
+* Add musicbrainz support: http://www.rupamsunyata.org/~decklin/musicbrainz-get-tracks
-* move the wav files to a different dir if the user wants to keep the tracks
* read the TOC and CDTEXT files and store them
-* read the complete CD to a single file, storing TOC seems a good idea to
- recreate the index entries if we decide to burn the cd.
-* read_and_encode function that uses named pipes for further ripping/encoding
- parallelization where possible
* read_and_encode_and_tag? :)
* Customizable post-read/encode/tag hooks
(http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=114851)
since abcde sees the encoding already finished. Same with tags and moves.
* Support UTF-8 tags
+
+* From slashdot: IMHO the most important aspect of an auto-ripper, is its
+ error-handling: what happens if a CD is too scratched to rip? How should it
+ react if someone tries to rip the exact same CD? make a new rip with another
+ name ? silently overwrite the old rip? etc.
+
+ Review the things that abcde does ;)
+
+* When cdparanoia encounters a problem (disk full, data track) when abcde
+ is run in batch mode, abcde still continues, but batch mode is disabled.
+ I.e. instead of normalizing all tracks together, they are normalized
+ separately, instead of running "lame --nogap" on all wavs, lame is
+ invoked once per wav.
+
+ Abcde should halt when a problem occurs, so it can be manually resolved.