Magnet links

[Stephen Michael Kellat]

On Tue, 28 May 2013 06:43:21 +1000
Jackson Doak <noskcaj at ubuntu.com> wrote:

> The Xubuntu website (and all of Ubuntu) currently only uses .torrent files
> for it's torrent downloads, this is an obsolete format, and nearly all
> other torrent sites (both legal only and sites that don't care) use magnet
> links, either solely or as well as .torrent files.
> 
> How they work:
> file: link to "trackers", then downloads from a list from the trackers.
> link: uses lots of confusing algorithms to ignore the need for trackers, so
> no "middle man" and nothing to host, no download as well.
> https://en.wikipedia.org/wiki/Magnet_URI_scheme has more info.
> both methods get similar amounts of peers
> 
> Magnet links are easier to use than .torrent files and it is very likely
> that if someone both has a torrent client and knows what xubuntu is, the
> know about using magnet links.
> The link may even be less confusing, because you don't download a file that
> some people may think is xubuntu.

Jackson,

You've made a proposition towards using magnet links.  This needs to be built upon, though.  How many clients are there in the Ubuntu repositories that handle .torrent files?  How many clients are there in the Ubuntu repositories that handle magnet links?  To make a switch we have to have sufficient clients able to handle such.  The command line tool in the package "aria2c" is the only one I have found in my cursory look tonight.

This is a useful thought exercise to go through.  A quick look at MirrorBrain.org shows that that magic redirection back-end software still doesn't have complete support for magnet links and it is what the openSUSE folks use to support a major chunk of their infrastructure.  It doesn't look like the Debian crew have adopted magnet links yet and a quick check at Mageia shows they've not adopted such either.

Just because it is old doesn't mean it doesn't work.  After all, the world continues to have emacs roaming around.  The two key things that would need to be addressed would be client availability and infrastructure/costs benefits that might be derived from use.

How would you address those two issues?

Stephen Michael Kellat