sabato 14 aprile 2012

TRIBLER and troubles


I'm writing this post in english because I think that could be interesting to have a feedback not only from Italy.
I tried TRIBLER a famous new bittorrent client, it is open source, it is different from the others bittorent clients because it is a peer-to-peer client, but maybe Wikipedia could help me to explain it better:
"Tribler is an open source peer-to-peer client with various features for watching videos online. The user interface of Tribler is very basic and focused on ease of use, instead of including features.[2] Tribler is based on the BitTorrent protocol and uses an overlay network for content searching.[3] Due to this overlay network Tribler does not require an external website or indexing service to discover content.[4] Tribler features include: video-only searching, experimental video streaming, and an integrated video player. Tribler is available for Linux, Windows and OS X."
So this is a good idea, but after installed it I saw a directory called:
/TriblerDownloads/collected_torrent_files
Inside it there are many .torrent files named with an alphanumeric long name [SHA1 code] (eg. 666a20xxxxxxxxxxa1d54816e4bxxxd42231xxx4.torrent) and extension .torrent, so they are really torrent files.
Why I have so many .torrent files into my directory?
I found this:
"A big feature of Tribler is remote search, for this to work Tribler clients are collecting .torrent files and downloading only those files from other peers. By having this cache of .torrent files, a Tribler client can provide search results to others as well as use these to show local results.
We collect the .torrent files beforehand, because we need them in order to start a download. If we would not collect them, then a situation could occur in which the download could not start because of the lack of the .torrent.

Most .torrents are roughly 15 KiBytes, thus we feel that caching at most 50.000 of them is not a big issue (translates to less than 1 Gig). The .torrent files are collected with a rate of 1 every second for discovered torrents in a channel.
Every 15 seconds, after connecting to a new random peers, another 10 torrents could be scheduled for collection.

If you want to look at your total number of collected .torrent files and number scheduled .torrent files, click on the Tribler logo. And look for the torrent collected and torrents in queue stats.
If you manually start removing the torrents from your collected_torrent folder, the first stats will not be correct anymore
."
So, if I'm using TRIBLER I'm collecting .torrent files from others peers, so I am "spreading" .torrent files of every type...they could be about pedo, copyrighted software/videos/music etc..
Yes, maybe it's possible to deactivate this feature, but how many people will have awareness of this?
The illegal .torrent files spreading could have legal consequences?
This is a big trouble in my humble opinion ; - )

UPDATES:
I received quickly a response from Niels Zeilemaker (TRIBLER developer) you can read it into the comments to this post.

# 1
We're slowly moving away from the .torrent collecting. Currently we have a background process collecting .torrents after receiving updates of a channel. This results in many .torrent which should be collected, only we have a ratelimiter which requests one .torrent every 1.5s.

As .torrent collecting is basically only performed on channels you have marked as favorite you will not automatically start collecting .torrents for pedo copyrighted software etc.
Di  Niels Zeilemaker  (inviato il 14/02/2012 @ 10:35:39)
# 2
Thanks a lot for your response, I hope your project will improve always better ;)
Good job!
Di  Nanni Bassetti  (inviato il 14/02/2012 @ 10:50:01)
# 3
Due to the design of ChannelCast we will still collect a small preview amount of .torrents.
Upon discovering a new channel (any type of channel) we will collect a snapshot of .torrents.

These can contain .torrents pointing towards copyrighted material. But .torrents themselves are not illegal, they only describe the data.
It's only a descriptor not part of the actual contents.
The Family Filter will only filter results in the gui.
Di  Niels Zeilemaker  (inviato il 14/02/2012 @ 11:50:48)