It seems very straightforward to add a Bittorrent client directly into a web browser. The OBJECT element in HTML provides enough rope to use Bittorrent. I'm imagining the most common use case to be embedding video into a web page.
We do need a new URI to invoke Bittorrent. I'm proposing btp1://URL#file, where "1" is the version number of the protocol to allow for future improvements, "URL" is the location of the .torrent file, with any pound signs and percent signs escaped (with percent hex) and "file" is the optional file within a multi-file torrent.
For truly huge torrents, recursive torrents may be specified with a btp1 URL for the inner URL, e.g., btp1://btp1://etc, with percent escaping as necessary.
Of course, the hard part is in the implementation in the browser. I imagine all the recent downloaded torrent contents are saved in the browser's LRU cache, and the torrents still in cache are all turned back on when the browser starts. The torrents get paused whenever the user is actively downloading something else.
It would be nice if the internet, both within your home and the internet as a whole, had voluntary Quality Of Service implemented, so you could mark your background torrents as lesser priority.
1 comment :
The previous version had the following text:
For truly huge torrents, a once-recursive torrent may be specified: btp1://URL##file, where the URL is for a torrent which contains a single torrent file inside it, and "file" is the file inside the second multi-file torrent. Other variations include btp1://URL## if both the first and second torrents are single file, and btp1://URL#file1#file2 if both are multi-file.
Clients may (but are not required to) permit further levels of recursion.
Post a Comment