Coral links prevent mefi-effect

August 28, 2004 10:29 PM

The future of the Internet? If your posted site is getting mefi'ed, you might consider also providing a link from Coral. Coral is a p2p network that replicates content across participating servers. Heres metafilter through Coral.
posted by vacapinta to MetaFilter-Related at 10:29 PM (9 comments total)

It's not really peer to peer, it's just distributed (they run a lot of servers in different places. From the clients point of view it's all still very much client-server). The archive.org wayback machine has been capable of doing something similar for a long time, though slightly less up to date as they don't work on-demand (the bright side being you can often access sites that are already down/gone). The wayback machine also scores points for rewriting links and images so that it works with absolute URLs too.
posted by fvw at 11:56 PM on August 28, 2004

Sounds like an original idea.
posted by Keyser Soze at 12:53 AM on August 29, 2004

it's a good alternative to the trusty ol' google cache anyway. thanks vacapinta.
posted by triv at 3:08 AM on August 29, 2004

Will this site allow reading of MetaFilter during its JRun errors? I take it the cache gets loaded as people request pages. So if enough of us casually read MeFi through this site will the cache be thus loaded for viewing when MeFi goes offline for a few hours? Or, is it so real time that the cache will get loaded with the JRun error page right away?
posted by caddis at 5:47 AM on August 29, 2004

Ugh, that's what metafilter looks like when you aren't logged in. I miss the image free days.
posted by Grod at 9:51 AM on August 29, 2004

I'm pretty sure it would cache the Jrun error, but then I don't know often they have it set to check for new versions of sites.
posted by reklaw at 11:33 AM on August 29, 2004

They check (at most) every five minutes (but if everything is ok, nobody will go through the cache and so it'll only attempt caching when it's dead. I don't know what HTTP code MeFi returns for JRun errors, but as long as it's not a 200 ("OK") the cache should realise it's not the proper data and not cache it.
posted by fvw at 3:21 PM on August 29, 2004

Well, it's not like we're going to have to wait long to find out.
posted by reklaw at 5:49 PM on August 29, 2004

We have found out - it will not work.
posted by caddis at 5:36 AM on August 31, 2004