BitTorrent is a tool for distributing files. It's extremely easy to use - downloads are started by clicking on hyperlinks.Whenever more than one person is downloading at once they send pieces of the file(s) to each other, thus relieving the central server's bandwidth burden. Even with many simultaneous downloads, the upload burden on the central server remains quite small, since each new downloader introduces new upload capacity.Windows web browser support is added by running an installer. A prebuilt one is available, but instructions for building it yourself are in BUILD.windows.txtInstructions for Unix installation are in INSTALL.unix.txtTo start hosting -1) start running a trackerFirst, you need a tracker. If you're on a dynamic IP or otherwise unreliable connection, you should find someone else's tracker and use that. Otherwise, follow the rest of this step.Trackers refer downloaders to each other. The load on the tracker is very small, so you only need one for all your files.To run a tracker, execute the command bttrack.py Here is an example -./bttrack.py --port 6969 --dfile dstate--dfile is where persistent information is kept on the tracker across invocations. It makes everything start working again immediately if you restart the tracker. A new one will be created if it doesn't exist already.The tracker must be on a net-addressible box, and you must know the ip number or dns name of it.The tracker outputs web logs to standard out. You can get information about the files it's currently serving by getting its index page. 2) create a metainfo file using btmakemetafile.pyTo generate a metainfo file, run the publish btmakemetafile and give it the file you want metainfo for and the url of the tracker./btmakemetafile.py myfile.ext http://my.tracker:6969/announceThis will generate a file called myfile.ext.torrentMake sure to include the port number in the tracker url if it isn't 80.This command may take a while to scan over the whole file hashing it.The /announce path is special and hard-coded into the tracker. Make sure to give the domain or ip your tracker is on instead of my.tracker.You can use either a dns name or an IP address in the tracker url.3) associate .torrent with application/x-bittorrent on your web serverThe way you do this is dependent on the particular web server you're using.You must have a web server which can serve ordinary static files and is addressable from the internet at large.4) put the newly made .torrent file on your web serverNote that the file name you choose on the server must end in .torrent, so it gets associated with the right mimetype.5) put up a static page which links to the location you uploaded to in step 4The file you uploaded in step 4 is linked to using an ordinary url.6) start a downloader as a resume on the complete fileYou have to run a downloader which already has the complete file, so new downloaders have a place to get it from. Here's an example -./btdownloadheadless.py --url http://my.server/myfile.torrent --saveas myfile.extMake sure the saveas argument points to the already complete file.If you're running the complete downloader on the same machine or LAN as the tracker, give a --ip parameter to the complete downloader. The --ip parameter can be either an IP address or DNS name.BitTorrent defaults to port 6881. If it can't use 6881, (probably because another download is happening) it tries 6882, then 6883, etc. It gives up after 6889.7) you're done!Now you just have to get people downloading! Refer them to the page you created in step 5.BitTorrent can also publish whole directories - simply point btmakemetafile.py at the directory with files in it, they'll be published as one unit. All files in subdirectories will be included, although files and directories named 'CVS' and 'core' are ignored.If you have any questions, try the web site or mailing list -http://bitconjurer.org/BitTorrent/http://groups.yahoo.com/group/BitTorrentYou can also often find me, Bram, in #bittorrent of irc.freenode.net