Deleting Content

I discovered this project this afternoon and I’m thinking through the process. As I read up on it, I see the word “permanent” a lot but as I read into it more it seems that if I’m running an IPFS node, I add something to my node, then I remove my node without anybody requesting my file, wouldn’t that make the file unavailable? Functionally that sounds like a delete process to me. Or even if somebody requests my files, if they GC without pinning my content then it would be gone forever.

In short, if content isn’t pinned on any server, wouldn’t GC eventually remove the file from the network?

Personally I dislike the fact that the IPFS website uses the term “permanent web” because it’s confusing. It’s using the term “permanent” because an IPFS link permanently references the content that it points to. This does not mean IPFS makes the content permanently available. As you point out, it’s up to you or the other people who care about that data to hold onto copies of the it. If nobody holds onto a copy, the data will go away. The link will still permanently point to it, but the data might not be there to retrieve.

The choice is between HTTP - a brittle system with volatile links vs. IPFS - a resilient system with permanent links. The use of content addressing does not make the web’s content permanent but it does make the links permanent while providing a plethora of powerful, dynamic ways to make data durable. Content addressing makes the system resilient, but not permanent. By contrast, HTTP links point to a location, which makes the system both brittle and volatile.

1 Like