Depending on the format of the archives (hopefully WARC), the owner could hand the entire archive over to the Internet Archive or Archive Team for ingestion by Wayback Machine.

If the concern is perpetual access to archived content but under your terms, that is where the cost comes in. Somebody somewhere is paying for power, cooling, connectivity, and disks. The Internet Archive estimates it costs them $2/GB to host data uploaded in perpetuity. Please consider donating if you're uploading content for permanent archival and/or deriving value from hosted content.

Pretty much. If you really want to maintain access to old stuff in perpetuity, you have to pay for it yourself by either storing it on your own equipment or paying someone to store it on theirs.

Sure -- but at least with something p2p hooked up to it I can let someone else benefit from my desire to assure my own access.

Yep, I agree.

The Internet Archive isn't perfect either. They can be pressured to remove pages.

It's not perfect. If you desire to store content that the Internet Archive must dark for whatever legal or compliance reasons, you'll have to cover the cost for that storage. The cost will always be non zero to perform such an operation.

nothing is permanent in life. heaps of torrents i want have 0 seeders.

ipfs's economically incentivised hosting layer (filecoin) offers storage in monthly increments, not perpetual.

> nothing is permanent in life

practically yes, but

https://github.com/philipl/pifs

An image of a Web asset is useful for archival purposes and may be indexed using text recognition ML.

Images take up way more space than HTML documents… actually, what with how bloated most “websites” are these days, that's probably not true any more.

Yes, the site owner's response seems excessively didactic when a system of mirrors would solve the problem.

> a system of mirrors would solve the problem.

A system of mirrors prevents a single node going down from taking the whole system (in theory at least, we've all seen plenty of times where failover goes poorly), but it doesn't do anything to ensure the long term survival of the system as people lose interest, lose the ability to participate, and sometimes die.

If you were around certain internet forums in the late '00s you might have run in to an image hosting platform called WaffleImages which was created in response to yet another popular free image hosting service locking down their embedding and ruining thousands of old posts. The goal was to distribute image hosting among community-operated mirrors, and it worked great for a few years. Over time though people lost interest while the rate of new mirrors getting added dropped to basically zero and eventually it fell apart.

Mirrors don't solve the problem, they move it.

How is a mirror not solving the problem of data loss if a single instance goes away?

The problem isn't a single instance going away. The problem is what happens when for whatever reason the owner stops maintaining the project. This is a common problem and despite all the bluster and buzzwords, the IT community hasn't really found a solution. Torrents are the only kind-of-solution, but they are not ideal for something that needs to be constantly updated.

Zeronet may help for content that need to be updated

> This is a very fair answer, and all "permalinks" are lies

Cool URIs don't change: https://www.w3.org/Provider/Style/URI.html

You can believe that cool URIs don't change or you could go the IPFS route. Similar to the way torrents have a 'health' score of plenty of seeders, and IPFS resources could live as long as people want that resource to exist (Not sure if that situation is baked into IPFS though).

Also: have you looked into Filecoin?[0]

[0] https://filecoin.io/