Some Steem dapp brainstorming..

SteemDNS - static page for search, pings steem custom json, forwards url to latest update to IP.
On the server side: Nodejs, or a custom script, listens for new IP from public IP lookup service and updates steem custom json.. can also be done manually.
so say I want to go-to the page for @sn0n/dev_machine
A static site, anyone can host their own copy, even locally, similar to Google in appearance, would look for a the most recent dev_machine custom json update under Hive account@sn0n and forward to the IP and port specified in a dev_machine ID.

OnChainJS - say Hive account@sn0n has a custom js library he wants to share. He can pop it under @sn0n/lib_name custom json, and you can load that JS code into the current running page with a custom loader bootstrapped from... With all the drama around NPM, seems a pretty cool idea....

Lots of other web ideas around using steem as a DB.... But then I wonder, can the public api nodes handle the strain if something suddenly gets hit with a digg/slashdot effect? I could include a type of load balancing mechanism in where a user of said scripts or libraries will load a list of known public api nodes, shuffle them, and rotate through the nodes as requests are sent... With respecting a users own personal whitelist and blacklist of trusted nodes?