3 ms·

This is probably the most well-thought-out comment here. It isn't a knee-jerk "it isn't permanent", "the tech won't work", "the funding won't work". I'm not loo

by hosh 7y ago

This is probably the most well-thought-out comment here. It isn't a knee-jerk "it isn't permanent", "the tech won't work", "the funding won't work". I'm not looking at this from the lens of an archivist, though I think that is important too. What do you think about the role of IPFS and FileCoin in digital preservation? Is there a role?

jrochkind1 7y ago
Thanks! I don't personally think very much of one, but i'm pretty down on blockchain in general. (Also not an expert at it, only have a surface-level IT guy understanding of it). I could see some role in using blockchain to keep track of _checksums_ of preserved files (nothing in the chain except the cryptographic digests of files stored elsewhere), as a way to have a hard-to-tamper-with multi-party-decentralized preserved historical record that a particular file was stored at a particular time, and has not changed since (whether from error or mal-intent). Just a record of what was stored when, not otherwise involved in the storage itself. If you want decentralized multi-party storage (which is cool, but I'm not sure it's necessary), there are protocols for that which don't require, or IMO have any reason to benefit from, blockchain technology. Again just my own not super well-informed opinion, and I'm hardly a blockchain expert (being pretty suspicious of it).
- hosh 7y ago
  Awesome, thank you for engaging :-) Let me see if my quick run-down here might spark something interesting. IPFS itself is not a blockchain. It is a piece of distributed infrastructure that allows people to share and store content in a distributed way. There are parts of it that are similar to the old torrent protocol. The main thing with IPFS is that the people who care about keeping a particular piece of data alive can "pin" it locally on their servers. Everyone accessing it will query other people who have pieces of that content -- just like with bittorrent. IPFS has no monetization or blockchain. FileCoin is the blockchain. It was intended to be the blockchain run on top of IPFS, although it may work on other content storage/transfer schemes. The idea with that blockchain is built around "proof of storage". That is, I can request someone store some bit of data for me, and you can offer to store it in exchange for something of value. Among one of the immediate things I wanted to apply IPFS to is this archive of digital material that contains a knowledge base for our current technology base. I have since lost track of that knowledge base. It was designed as a CD you can burn, and get all sorts of information. The idea was to preserve this information in case of something catastrophic happening to our civilization. One thing about IPFS, and perhaps FileCoin, is that it allows for decentralized knowledge bases. Stuff that could be administrated at a local, community level, and yet federated with many communities. An example: gardening. With the pandemic, there had been a lot of people interested in gardening for the first time. However, how you plant, what you plant, how you manage those things are depended upon local conditions. I live in the low desert, and what I need to do for here is different from my friend who live in the high desert, or in the SE US.
  - jrochkind1 7y ago
    Yeah that all makes sense actually, thanks. I think the level of technical sophistication of most cultural heritage organizations is such that I don't entirely trust them to get it right with tried-and-true non-novel technology... if I were in charge of relevant efforts (which I so so am not), I'd want to get some basics in place (including clear business models) before doing more than "side project leaning experiment" with such things. And then add in the novel technologies little by little. One distributed-storage infrastructure that has some uptake in cultural heritage preservation, built on probably much less elegant and exciting protocols, is https://en.wikipedia.org/wiki/LOCKSS https://en.wikipedia.org/wiki/LOCKSS
    - hosh 7y ago
      That's a great point about user/organizational sophisitication. Hmm. I'll check out LOCKSS.

by hosh 7y ago

jrochkind1 7y ago

Thanks! I don't personally think very much of one, but i'm pretty down on blockchain in general. (Also not an expert at it, only have a surface-level IT guy understanding of it). I could see some role in using blockchain to keep track of _checksums_ of preserved files (nothing in the chain except the cryptographic digests of files stored elsewhere), as a way to have a hard-to-tamper-with multi-party-decentralized preserved historical record that a particular file was stored at a particular time, and has not changed since (whether from error or mal-intent). Just a record of what was stored when, not otherwise involved in the storage itself. If you want decentralized multi-party storage (which is cool, but I'm not sure it's necessary), there are protocols for that which don't require, or IMO have any reason to benefit from, blockchain technology. Again just my own not super well-informed opinion, and I'm hardly a blockchain expert (being pretty suspicious of it).

hosh 7y ago
Awesome, thank you for engaging :-) Let me see if my quick run-down here might spark something interesting. IPFS itself is not a blockchain. It is a piece of distributed infrastructure that allows people to share and store content in a distributed way. There are parts of it that are similar to the old torrent protocol. The main thing with IPFS is that the people who care about keeping a particular piece of data alive can "pin" it locally on their servers. Everyone accessing it will query other people who have pieces of that content -- just like with bittorrent. IPFS has no monetization or blockchain. FileCoin is the blockchain. It was intended to be the blockchain run on top of IPFS, although it may work on other content storage/transfer schemes. The idea with that blockchain is built around "proof of storage". That is, I can request someone store some bit of data for me, and you can offer to store it in exchange for something of value. Among one of the immediate things I wanted to apply IPFS to is this archive of digital material that contains a knowledge base for our current technology base. I have since lost track of that knowledge base. It was designed as a CD you can burn, and get all sorts of information. The idea was to preserve this information in case of something catastrophic happening to our civilization. One thing about IPFS, and perhaps FileCoin, is that it allows for decentralized knowledge bases. Stuff that could be administrated at a local, community level, and yet federated with many communities. An example: gardening. With the pandemic, there had been a lot of people interested in gardening for the first time. However, how you plant, what you plant, how you manage those things are depended upon local conditions. I live in the low desert, and what I need to do for here is different from my friend who live in the high desert, or in the SE US.
- jrochkind1 7y ago
  Yeah that all makes sense actually, thanks. I think the level of technical sophistication of most cultural heritage organizations is such that I don't entirely trust them to get it right with tried-and-true non-novel technology... if I were in charge of relevant efforts (which I so so am not), I'd want to get some basics in place (including clear business models) before doing more than "side project leaning experiment" with such things. And then add in the novel technologies little by little. One distributed-storage infrastructure that has some uptake in cultural heritage preservation, built on probably much less elegant and exciting protocols, is https://en.wikipedia.org/wiki/LOCKSS https://en.wikipedia.org/wiki/LOCKSS
  - hosh 7y ago
    That's a great point about user/organizational sophisitication. Hmm. I'll check out LOCKSS.