Critical to making video a first class citizen on the web is extending the properties of other first class citizens like text and images as they apply to video. These properties should include:
The xiph/annodex set of technologies are well suited for archiving these goals. Silvia's position papers go into more detail about the xiph/annodex technologies set.
Critical to strong user experience is powerful standards based api for web services to build on. The whatwg group has made good progress on defining this api. http://www.whatwg.org/specs/web-apps/current-work/#video
The metavid project has modeled an user experience with online web video using free and open source xiph/annodex technologies set. Metavid has been archiving public domain house and senate footage and making it available online for close to two years. All the video is encoded with ogg theora video codec. All the metadata is encoded to CMML an xml language for continues video markup. Arbitrary segments of the video streams are available for download or in-browser playback via mod_annodex a server side script for dynamically splitting ogg theora clips. Client side playback is handled via oggplay, java crotado, mplayer plugin, vlc plugin or natively supported by the soon to be released Firefox3 and Opera web browser. Scripts such as mv_embed transparently handle the selection of whatever client side playback method is available.
Metavid is close to releasing a mediaWiki extension that will allow groups of participants to collaborate in creating close captions and metadata annotations for video streams. These captions will be made available via CMML feeds and or be serialized into the ogg stream via annodex. They will serve both as a search index into the video and text for the hearing impaired. Web interfaces can make this CMML text content visible.
Additionally the metavid Wiki extension includes basic video sequencing/editing. Critical to decentralized online video editors will be wide support for a common transclusion video url request system. This allows online services to reference video segments across distinct web services and hosting environments while preserving the associative metadata.
The metavid project creates one possible interface for collaborative video usage online enabled by and implemented as a FOSS system.
The existing http protocol is functional for video distribution and fits into the exiting web infrastructure. Most video content should be downloadable and unencumbered by proprietary container or formats. This will help facilitate mashups, reuse and fully realize the potential of online video. Critical to wide adoption of any proposed standards is inclusion of free software implementations. Specifically the audio/video codec and container must be freely implementable. The should W3C seek to standardize around the best open formats available at the time of standardization. The xiph ogg container and set of audio/video codecs are ideal.
I could demo portions of the metavid interface highlighting the annodex technologies at work in the metavid site.