HTTP as a file system

I want to be able to treat URIs as files in unix:

$ cat /http/

There are all sorts of implications here. In Unix, you can read and write files. You can also **execute** files:

$ /http/some_music_site my_mp3_file.mp3

Arguments could be passed as byte streams or whatever is simplest (or most appropriate for the data type). Text should be preferred — the receiving file (program, executable, web app?) is responsible for receiving data elegantly just like any other unix program would.

The concept of, say, a blog, could be transformed. I’m currently stuck with the wordpress editor if I’m using a wordpress blog. I could edit a blog post using whatever editor I wanted, as long as my blog was implemented as an HTTP API that understood how to be treated like a unix file.

This could be a library that HTTP servers could use, or it could just be a convention. It would be awesome to be able to do:

$ mkdir


Zero Editorial

I recently clicked a link on Hacker News entitled “There’s a new open source cloud in town. Meet Apache CloudStack”. It was a link to GigaOM **about** Cloudstack.

So then I did a google search for “apache cloudstack”:


Not a single non-blogspam link. It turns out, I wanted []( Who knew.

I’m going to build a site where anybody can contribute links, with the only requirement that it cannot be a link to something “about” what the really interesting stuff is. No editorializing. “Apache Cloudstack” will link to the cloudstack website. A link about a new apple product will link to Apple’s video, or to the product page.