Automatic addition of trackers
With the probable closure of The Pirate Bay (TPB) in the near future the torrent scene will be more decentralized. Especially if the TPB tracker goes down. Luckily there are new upcoming trackers to take its place. The problem is that they are not in many of the torrent files. Besides adding the trackers manually there are some ways to fix it. One way is to re-upload all torrents with more trackers. That doesn’t seem like a plausible method. Another way is the have an indexer site (like TPB, Mininova or BTjunkie) update all the torrents with more trackers. This doesn’t seem like a bad idea.
Another idea, and this is my idea for the client, is to have an option in the client to automatically add more trackers to all torrents loaded. That would be very useful for torrents that do not include any new public tracker such as Openbittorrent (OBT)1 or Publicbittorrent (PBT)2. There could be an editable list of trackers to be automatically added, or there could be function that shares all public trackers for all torrents though I’m unsure of what the consequences will be of that. Even more, one could have peers share public trackers. In that way any new public tracker would quickly spread. There is reason to be cautious, however, a third party, say RIAA, may set up such a public tracker just to sniff IPs.
Automatic force-seeding of torrents with no seeds
The goal is to keep rarer torrents alive easier. Currently I have to manually look through my torrent client for torrents that currently have no seeds and then force seed them so that the peers can finish their download.3 It should probably force seed for a set time. If it forced-seeded only until there were at least two seeds, then complications arise. Imagine that there are two people who are the only seeds and both their torrent clients have this feature. Both their clients would then force seed the torrent until they spot another seed, but then they can see each other and so they will keep bouncing on and off force seeding while the peers would not get much data. A mix of these two ideas may be a good idea.
Footnotes
3We’re presuming here for simplicity that if there is no seed, then there is not a complete file in the swarm. This is false in some cases.
