Hacker News Reader

Comment by JumpCrisscross | Hacker News Reader

> What happens when you leave that company? New company has to go through the same laborious process.

I like how de-centralisation is a pro until it's versus centralization on a blockchain.

TL; DR If everyone can agree on a single blockchain, everyone can agree on a single data format. That data format agreement is faster, more adaptable, more extensible and more resilient (in not requiring users to keep private keys) than a blockchain.

Screaming blockchain at everything is a new twist on an old mistake: thinking political problems have technical solutions.

> hate if you will

Discarding reasonable criticism as hate is immature. So is arguing "if you think X is not important, you're fooling yourself."

View on HN

> TL; DR If everyone can agree on a single blockchain, everyone can agree on a single data format. That data format agreement is faster, more adaptable, more extensible and more resilient (in not requiring users to keep private keys) than a blockchain.

Exactly. If you're looking for a model of personal medical data ownership, DNS would be better.

Various companies that hold your data (registars) and allow you to share with the appropriate parties (zone files), or delegate that ability if you so choose (name servers). If you want to transfer that data to another company, you can (domain transfer).

Furthermore, I question the assumption that immutability is entirely useful for medical records.

Immutability is useful for medical records for accountability purposes, to prevent doctoring (no pun intended).

What if the data is wrong?

You append a correction, it's the same process as an edit.

So, use git?

What if say you bribe someone who can somehow get access to the files and modify them?

What if say you bribe someone with access to the private keys to append a false entry?

Indeed, you must have a Chain of Command/Chain of Trust that the majority of people will trust. I believe there are ways.

So what's the advantage over using git?

1 extra level of security? Perhaps the added cost of the system doesn't warrant it, perhaps there are better ways to mitigate.

I'm not an expert in the area and don't know the nuances as much as I'd like, like I've wondered if there's a way to have 2 keys - so in the case of blackmail, you give a different key than normal, which will work - however then there's a way to trace.

Obviously the best system is having a system of governance that is stable and accountable. The situation to perhaps design for is say tyrannical dictatorships that may rise.

I suppose the most important aspect of git would be redundancy of data, along with perhaps immutable and/or offline archives; evidence may be changed without anyone ever knowing though to check, unless perhaps there were regular (every 5 years?) consistency checks between archives and present-day/online historical data?

If you start from a place of distrusting your medical professional, now you need a way of uniquely identifying them in a way that can tie back to them for accountability purposes. So we've introduced some kind of notarized identity service.

Furthermore, doctors are only going to read the most recent entry for any given data point: doctors won't read the entire file, especially in emergency situations.

Just because the status quo and doctor's behaviour is currently a certain way, doesn't mean it's the right way. The efficiency lacking in our health-"care" systems is astonishing - there needs to be self-regulation mechanisms and accountability.

Emergency situations is an orange if everything else is an apple, so of course there can be protocol and acceptable boundaries for each.

Humans are humans, mistakes can happen - especially that we aren't aware we're making, and therefore bound to make them again. Accountability will lead to necessary learning and improvement.

Reply on news.ycombinator.com