2 ms·
Thank you for that! This was particularly interesting: Certain versions of Unix, and early network file systems, used paths of the form //hostname/path to acce
by keepamovin 2y ago
Thank you for that! This was particularly interesting:
Certain versions of Unix, and early network file systems, used paths of the form //hostname/path to access 'path' on server 'hostname'.
It directly ties in with the URL format:
scheme://hostname/path/to/file
How cool it would be if we could go:
cd //google.com/search?q=quack+like+duck
and we would get a directory with 1 file per result (or to be really unix like: 1 directory per result, with a set of files and directories for different aspects of the result hehehe :))
- wruza 2y agoThis is an old idea behind “everything is file” and plan9, afaik. FUSE allows this, you can easily write a “driver” that does exactly this on a mountpoint like /mnt/mywebfs/whatever.com/search?q=…. /proc (procfs) is basically this but for kernel structures.
- keepamovin 2y agoI recall somebody did this with www, I think at first using a browser extension? Webfs possibly. It is a really cool idea. I find it makes many things easy to deal with, and those that aren’t? Interesting to think about. The hierarchical abstraction of a file system is a good model from many things