Notice that the read link is only offered on the XML files. This is because libferris tells the client that those are not natively "directories" but can be seen that way if you like.
I then simply click down to the "barry" XML element in the mounted XML file. Editing the barry entry will write the data back to the server asynchronously. The terminal is used to verify that things went on back.
The second browser tab shows a mounted PostgreSQL table which has ID and Message columns. If I select edit on the message column I get to update the tuple as desired. Notice that the data grid in the browser is updated when the data is saved on back by watching the "message" column in the browser. I then use psql to verify the update from the command line.
Other interesting possibilities include mounted log files being split into columns in the web view, or grabbing some data from plasma on the server, streaming data from gstreamer or zoneminder, or anything else that ferris can do as a VFS. And as I tend to want ferris to mount it all^TM, the sky should be the limit! Heh, so I got the catchphrase in there emes.