Video is accessible when every person, no matter what limitations in language understanding, hearing, seeing, or other senses, can follow what is happening in a video and navigate it. Video accessibility is fundamentally about providing textual and other additional information about the video to help provide information in channels other than eyes and ears.
Captions and subtitles are only one type of accessibility features - there are also audio annotations for the blind, and many other text representations that are related. For years, people have been requesting a solution for Ogg content with subtitles/captions. So far, the main solution was to create a text file (e.g. a srt file) and load it together with the video file into a media player that was then able to do the subtitling ("soft subs"). Now that Firefox supports Ogg Theora/Vorbis out of the box, an encapsulated solution is required ("hard subs").
Silvia is working for Xiph and Mozilla on this and has recently proposed a generic mapping of "text codecs" into Ogg. This will encapsulate the W3C TimedText standard as well as your fansubber's typical formats.