I disagree that metalinks are "better" than torrents. Metalinks may be 
standardized but bittorrent clients do not care about them and most people 
don't have a program that can download metalinks (at least in 2011), 
whereas almost everyone (in my experience) has a bittorrent client. The 
good news is that the .torrent file format already supports "web seeds" 
(since a few years ago) which as the name suggests allow you to put an HTTP 
URI inside the .torrent file as a web-based seed which a torrent client can 
fall back to when there are no other seeds on the torrent. We should offer 
.torrent files with our HTTP download locations listed in the .torrent file 
as web seeds. Metalink files should be an alternative for those who 
actually have a metalink client. I know that I for one was pretty surprised 
to find metalinks but no torrents on sagemath.org when I first downloaded 
Sage. I ended up just downloading by HTTP since I couldn't be bothered to 
install aria2.

There are also other ways that you can more intelligently do automated 
seeding of torrents, by the way; you could just run a daemon on 
boxen.math.washington.edu which connects to our torrents' swarms and 
throttles uploading if there are enough other seeds around. This will 
mitigate the possibility of the whole torrent just turning upon boxen and 
downloading from it, which offers us no bandwidth savings over just the 
HTTP downloads (though this is unlikely to happen anyway when using the 
.torrent file "web seed" field instead of installing a bittorrent client on 
boxen, assuming users' bittorrent clients are well-designed).

-Keshav

----
Join us in #sagemath on irc.freenode.net !

-- 
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to 
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to