[lmilcin]

It is a hash of a block, but not a chain.

For this to be a chain, it would have to be published multiple times, each time containing references to previous blocks under the published hash.

Other than that I am pretty sure we can find many more examples. Hashing and publishing the hash is pretty popular way to prove timestamp of a secret without revealing the secret, yet. I did it myself a couple of times when I wanted to be able to prove later that I had a particular information before some specific time.

[sorokod]

Other than that I am pretty sure we can find many more examples.

This statement could be made much more interesting and convincing if you provided some actual examples.

[sjs382]

It's fairly common (at least among the security community) to post a hash of a claim to Twitter, that they can reference in the future to prove that they had X information on Y date without disclosing the actual information until later.

For example, I could post the hash "9712f2488fa00fc5b11b657995ac10c9" to Twitter and then after the prediction comes to pass, I can reveal that the hash was for the phrase "Pittsburgh Steelers will Defeat New Orleans Saints 30-21 in Super Bowl LIII", proving that I had some knowledge or prediction that I didn't want to reveal at the time of the prediction.

[blocked_again]

Wait but this is not fail proof right. For example I can post that Kolkata Knight riders will defeat Mumbai Indians along with a post that Mumbai Indians will defeat Kolkata Knight riders and delete the wrong twitter post once the competetion is over?

[logfromblammo]

Presumably, one can assume that once it is published publicly, someone else could make a note of it, and tattle on the author whenever they delete a hash without revealing the source of it.

In practice, nobody is attending to anyone that closely, aside from a handful of public figures, and shenanigans like those you described might actually work to get someone more attention, only to fail once they have enough. That's why a publication mechanism that disallows deletions is desirable for such hashes.

[sorokod]

How is this similar to:

"Instead of posting customer hashes to a public digital ledger, Surety creates a unique hash value of all the new seals added to the database each week and publishes this hash value in the New York Times. The hash is placed in a small ad in the Times classified section under the heading “Notices & Lost and Found” and has appeared once a week since 1995." ?

[ChrisClark]

How is it similar? I think we could rephrase it.

- Post a hash of a document to Twitter.

- Put a hash of a document in an ad in the NYT.

Quite similar.

[lmilcin]

It's not difficult to find examples:

https://twitter.com/gruber/status/991122089557004288?lang=en

There is even whole industry around the concept, it is called trusted timestamping: https://en.wikipedia.org/wiki/Trusted_timestamping

It basically works like this, you compute a hash of the document and send to some kind of service that will either publish your hash with the timestamp or will also sign the hash and the timestamp cryptographically without ever touching the document.

Later, you can refer somebody to the external service or let them examine the signature and the hash to prove that you were in possession of the information before certain date and time.

[JWett]

"Hash and Sign" timestamping is flawed in that it relies on the trust of a central authority. An insider with access to the private key could easily backdate a timestamp and sign any document/timestamp combination. Surety's widely-witnessed approach was meant to fix this issue.

Chaining the hash values made it fundamentally impossible for any malicious actor to generate a notary certificate for a future document that would roll up and produce the correct super hash value that is woven into the chain.

[phyzome]

It's a chain: https://news.ycombinator.com/item?id=17865168

[lmilcin]

Thank you for posting this and clearing it up. This was missing from the original article.

[phyzome]

Yes, I found it quite frustrating as well. -.-

[FabHK]

It's a block chain. From the article:

> a copy of that seal and every other seal created by Surety’s customers is sent to the AbsoluteProof “universal registry database,” which is a “hash-chain” composed entirely of Surety customer seals. This creates an immutable record of all the Surety seals ever produced, so that it is impossible for the company or any malicious actor to modify a seal.

Another well known example of course is git (which slightly extends the definition to a directed acyclic graph).