Archive.org is a great project that one can't recommend highly enough.
It's not "just" collecting huge amounts of data, it is a living archive. For example, when they publish old Amiga software, they publish it ready to use. And I don't just mean polished disk image files. I mean ready to use right in the browser. Click on a ancient game, and play it immediately in your browser!
That's exactly the kind of attitude towards archiving that we need. Like one of the more modern museums where you are allowed to touch (and interact with!) things.
I don't think IPFS and the like can solve the archival problem. They rely on the willingness of lots of people to donate lots of storage space indefinitely.
I don’t get why people think IPFS is only a volunteer thing. The same way bitcoin quickly evolved from people mining on laptops to professionally run datacenters, so to will providing resources for IPFS.
Filecoin will allow people to monetize making content on IPFS available, with the ability to prove the content meets certain availability guarantees via smart contracts.
Making content available via IPFS will become the new way to mine cryptocurrency.
But there are other protocols incoming too, like Swarm for Ethereum, Storj, etc. Some of them already available just like IPFS.
Not an apples-to-apples comparison—IPFS is the only one that’s an offline first, peer-to-peer, distributed versioned file system. Swarm is interesting but it’s for small storage for smart contracts; it’s not a general purpose, low-cost storage option for storing terabytes of data. You’re not going to take a snapshot of Wikipedia on it, for example: https://ipfs.io/blog/24-uncensorable-wikipedia/
I doubt IPFS will automatically win this because it was the first to make a glorified BitTorrent client available via HTTP.
This statement doesn’t make sense—IPFS is designed to replace HTTP, not run on top of it. While it shares some similarities to BitTorrent like using a DHT for content addressing, it’s really a different thing.
>Swarm is interesting but it’s for small storage for smart contracts; it’s not a general purpose, low-cost storage option for storing terabytes of data.
Swarm is intended to do exactly that.
IPFS, atm, does not store terabytes of data. If I were to just dump in my data, it would be unusable within the hour as there is no incentive for nodes to keep those terabytes active and around.
Swarm already hosts static websites, snapshots of wikipedia are feasible.
>IPFS is designed to replace HTTP, not run on top of it. While it shares some similarities to BitTorrent like using a DHT for content addressing, it’s really a different thing.
IPFS does not address dynamic content properly, IPNS is way to slow to allow websites on the scale of google to operate sensible. I doubt IPNS could handle a decently sized subreddit in terms of activity.
IPFS is unlikely to replace HTTP since both protocols address different problems. However, as it is usable today, IPFS is little more than a cache that can store some data for a bit until nobody is interested in it.
Last I recall, Filecoin is not IPFS, it merely works on top of IPFS. You may revie your argument and replace every occurence of "IPFS" with "Filecoin", in which case it would still compete with Swarm, Storj, etc.
We need to have a well-funded, well-run, serious effort at preserving digital history.