I essentially took this configuration from here and modified it.

Otherwise read the man pages for tmux as it is quite good. Just remember that I rebound the prefix key from C-b to `.

<script src=”https://gist.github.com/762305.js”> </script>

I decided that mutt was worth a try. So I put together a configuration that works with Gmail by scouring the internet. First follow the directions listed on Lifehacker’s article but use this for the configuration. Change my email address to your’s at the appropriate place.

Also, I want mutt to ask me for my password everytime I use it. This is because… ARE YOU FUCKING CRAZY FOR STORING YOUR PASSWORD IN CLEARTEXT?

When ever you delete from mutt what it does is removes the “label” so stuff is never deleted but moved to “All mail”. So in this configration there is no way to delete mail. If you want that please look at the Lifehacker configuration.

<script src=”https://gist.github.com/762304.js”></script>

To make life easier I migrated this code by Jack Moffitt for my own use. I didn’t need anything fancy like draft management since I use org-mode for that. But it allows me to create a new post from a keystroke.

<script src=”https://gist.github.com/762301.js”></script>

In the context of a torrent indexer how do you trust that the person who is pushing you torrents is not pushing you crap. There is a level of concern here. Trust is an important aspect of gaining new torrents, but at the same time we don’t want to trust someone too much such that they will start pedling junk to us.

I am still trying to think about the solution to this problem. However, I think the PGP Web of Trust model is quite good. Trust someone when you know that others have trusted them as well. However, this also posses as the weakest link. Do we really trust to be smart about crypto security? Unfortunately, I don&rsquo;t think we can (and this is speaking from experience…).

The second part is how to deal with rogue indexers. Sites who have decided to go bad pushing junk. This part should be a bit more straightforward. Just make sure that a certain level of torrents from that indexer are verified or checked. If enough people mark a torrent as bad and further if enough torrents coming from a certain site are marked as bad then we know that it’s a rogue torrent.

In this situation all we have to do is block that Indexer from pushing us torrents, mark that public key as untrusted and push to all the indexers which are following that that indexer is a rogue indexer. This way the rest of the network is aware of such rogue nodes.

The proposal is to use the latest web techniques such as WebHooks and Oauth to create a distributed indexing infrastructure for torrents. Furthermore, a site operator must have the ability to maintain what torrents appear on the site (maintaining quality) and further who he is going to trust to push files to his server. Since we cannot trust any single centralized authority we must have a decentralized trust model. This is why PGP will be used as a mechanism for maintaining the Web of Trust between index site maintainers.

I am currently writing this as a Ruby on Rails applications. It is available here. Right now I have a basic Rails app running where a person is able to upload files and the tracker part is near completion. However, my main interest is the Indexer to Indexer communication. I am currently trying to figure out how to do that correctly. Since I have no crypto experience I am trying to get myself familiar with crypto concepts before making a move. However, the main problem that I will have going forward is that the PGP library that I would like to use for the Web of Trust stuff is immature. I will likely have to help build that up before I can continue. Overall, this is an interesting project for me right now and I hope to build it up to something which is quite durable.