I think the idea behind content addressable databases like this is that they are supposed to be permanent. Putting it out into space may be a way of saying "its never going down".
"there is no GUARANTEE about file preservation there at all (unless you pay to people hosting your files specifically)"
Sure your files are accidentally still present in the cache. Good for you. The problem is deceiving other people who may think it's somehow spacemagically-interplanetary guaranteed.
The only thing I can find from official resources is documentation saying IPFS does not do that at all. Example from https://docs.ipfs.tech/concepts/what-is-ipfs/#what-ipfs-isn-...
> What IPFS isn't
> IPFS is not ... A storage provider: While there are storage providers built with IPFS support (typically known as pinning services), IPFS itself is a protocol, not a provider
Reading your link I don't quite understand how come they claim IPFS is both a protocol AND a network of nodes AND not a storage. For example TCP is a protocol, but TCP is not a network of nodes even if there is a network of nodes using TCP to communicate.
I think this is some cheap wordplay to play pretend distributed filestorage and but also absolve themselves from any responsibility in case something goes wrong "IPFS is only a protocol, you probably was holding it wrong".
Sounds like those people either don't understand what they're saying, or they're mindlessly shilling for Filecoin. Both are obviously shitty, but I don't think it's fair to point the team behind IPFS in a bad light because of what others say about it.
> I don't quite understand how come they claim IPFS is both a protocol AND a network of nodes AND not a storage
IPFS is a protocol, IPFS is also the name of the network. IPFS doesn't automatically distribute data, hence it's not "storage", but a protocol for doing storage.
In the analogy to TCP, TCP/http are the most common protocols, and the most common network is what we call "Internet". Initially, there was many competing networks, until eventually The Internet overtook all the other networks and essentially "won".
(edit: again, gross over-simplification about the history the internet and more, hopefully won't be too misleading. Better overview can be found here: https://en.wikipedia.org/wiki/History_of_the_Internet)
ipfs is not a fileserver. but you can use it to host files.
Start thinking that way and it will take you to a whole new universe.
It doesn't guarantee that the content the ID resolves to is available, but once you have a content-addressable ID, you can be sure you'll be able to get the data as long as the data is available somewhere on the network.