5))The caller_pc sends to the responder_pc the string ‘picture_jhon_2009.jpg’ over UDP.
Here we assume the caller_pc knows the responder_ip;
6))The responder_pc replies to the caller_pc with the string ‘QmRBkKi1PnthqaBaiZnXML6fH6PNqCFdpcBxGYXoUQfp34a’, over UDP.
Here we assume the responder_pc knows the caller_ip;
7))The caller_pc download the file QmRBkKi1PnthqaBaiZnXML6fH6PNqCFdpcBxGYXoUQfp34a, with ipfs_ipget
In the #smalldatabase, the user can change the old entry with the new entry:
OLD ENTRY: ‘picture_jhon_2009.jpg’, QmRBkKi1PnthqaBaiZnXML6fH6PNqCFdpcBxGYXoUQfp34a
NEW ENTRY: ‘picture_jhon_2009.jpg’, QmRBkKi1PnthqaBaiZnXML6fH6PNqCFdpcBxGYXoUQfp35d
Content Based Addressing is a feature of IPFS, it shouldn’t be considered as a limitation.
Your idea is nothing wrong, but this is against the philosophy of IPFS. You can have dynamic content based on IPNS, where your IPNS can point to any hash, or possibly other IPNS address. But having lot of IPNS addresses is not good either, IPFS is meant to be static, so that the files can be cached and later sent to other nodes. I think you should at a look at Decentralized Databases like OrbitDB and then have dynamic content over it.
The script is still the same, the script still has same IPFS hash. So in your case you are not actually creating any dynamic website at all, the way you get the IPFS hash is independent process from getting the content with IPFS.