The 708 decoder has been a work in progress for a long time, 
but it's never been finished simply because there's no 
pressure (all broadcasts come with 608 captions). Anyway it's
something that we want to complete sooner than later, so each
new CCExtractor release will at least take care of a couple
of items in this to-do list.

The people at narlak.com were nice enough to generate a test
708 stream for CCExtractor, which helped quite a bit with the
initial stuff.

Current status is that a (poor) .srt file can already be 
generated, so should a broadcast with 708-only captions appear
a transcript would be available.

To do:
- Timing: For each .srt line, it's off by a second or so. 
- Colors: Currently unsupported.
- Pen sizes (useless in .srt of course, but the decoder should
  handle them internally).
- Roll up captions.
- SAMI output.
- Move everything to a 16 bit character set, to support 
  non-Latin character sets.
- Someone (from Korea, I think), reported that they have 
  708-only captions over there, in Korean. While I don't speak
  Korean I'll be happy to work on it a bit and see if I can make
  CCExtractor fully support Asian languages. I do need samples
  though. No samples, no support.
- A few commands are not yet supported, specifically those related
  to delay. 
- Detect and extract captions from MP4 (MOV) files, handled by gpacmp4

Done (18.08.2015):

- Major refactoring
- Librarized as much as possible (global vars moved to dtvcc context)
- Added some control commands support
- 16 bit charset support added (not "moved everything to 16-bit", but they could be successfully exported)
- SAMI output added
- Transcript output added
- Timed transcript output added
- Added colour support (only one colour/style can be applied for the whole screen at the moment)
- Roll up captions handling