Briar and Bramble: A Vision for Decentralized Infrastructure | Hacker News Reader

Briar and Bramble: A Vision for Decentralized Infrastructure

dymaxion.org

113 pointsericdanielski11 comments

This essay is and outstanding statement about decentralization and why it's important. It connects so deep into me that I feel it in my bones. This is the technological infrastructure we need, going to give there codebase a thorough inspection and see what I can build with it. Also, I love the finnish people, shoutout to my previous employer Reaktor Oy who did the honors of introducing me to the brilliance of the finnish people and culture.

While I love decentralization progress, the great majority of end-users look for UX don't matter what is behind. They don't care if it is decentralized, federated, or centralized. Git is decentralized but most people use GitHub and the UI cannot be decentralized. Also there is a lot of irony here where most decentralized projects use Slack or Telegram for messaging.

On mobile, offline first products simply work better and provide a better user experience. That doesn't require decentralization, and there are some apps were it won't be useful.

Totally with you on the disappointed part. People want things to work, but they don't care why something works and something else doesn't. So usually they allow others to bullshit them into using mediocre solutions just because these were cheaper to implement for the "provider".

However decentralized systems have huge advantages. E.g. in an unsure system (i.e. the normal broken unreliable network, especially when on mobile) decentralization helps to keep the few things running that by coincidence currently work. In a centralized system you always have a few things that really, really need to run, otherwise nothing works. That's for instance why the "ugly" email has survived the competition against "userfriendly" chat apps like ICQ and QQ. People still use email today. So even without much interest in them there can be quite some value for people who implement and design them.

I looked into Secure Scuttlebutt community, and while I didn't muster up the courage to really participate, I kinda like that they take dogfooding seriously.

Their protocol started as a secure peer-to-peer log-store, with first application as a white-list based social-network [1]

The more interesting part is a git plugin that allows to store repositories in the distributed log as well. This is complete with a nice web-UI and some rudimentary issue management [2]

If you had right group of people, you could use this on a local-network, with no internet access, and you'd have reasonably nice environment for chatting and collaborating :-) In practice you'd want to be on the internet and connected to at least one scuttlebut pub to actually see other people, unless you really organized some sort of offline-cabin-retreat-hackathon :-)

[1] https://www.scuttlebutt.nz/ [2] https://git.scuttlebot.io/%25RPKzL382v2fAia5HuDNHD5kkFdlP7bG...

That's why it's important for us who are working on decentralized projects to not use "decentralized" itself as a feature but rather use the side-effects of what decentralized technology brings.

The first ones that comes to mind is:

offline-first, local-second and internet-third. Basically, the applications you use day-to-day should work no matter how you are connected. If you are on a local lan, the application should work with the rest of the clients who are on the local lan. Slack would be amazing if when it looses internet connectivity, it simply tries to connect directly to other clients around it.

Performance. Decentralized applications can be much faster than centralized ones, for the majority of the people on the internet. If I'm in South Africa and using a web application with it's servers in US, there is a maximum speed the application can work at, no matter if I'm working with someone next to me, or in China. With decentralization, we can move logic closer to the people needing it.

Security. Less important for the average person but more and more people are caring about it. Apple seems to have a focus on it that actually makes security-conscious people chose them for it. Not having everything going through the same centralized service makes that centralized service less of a target, and decentralized technology usually does a way better job on using encrypted communication because it's needed when you send data peer-to-peer.

If instead of applications writing "It's decentralized!" on their landing pages like that's actually a feature, they explain the benefits of decentralization, I think it becomes a lot more interesting for the average person.

Disclaimer: I work on IPFS, a P2P protocol.

Its been a long time since Ive seen such a spread in technological capability and economic practicality.

Decentralization has all these fascinating bits and pieces but not a single consumer will care unless the consumer is another dev. Id argue blockchain falls into the same category. What we need are folks who can build products whose economics need these tools without having to promote the implementation as the differentiating factor.

IOW all this stuff will succeed when we no longer have to mention the word “distributed”.

This protocol seems pretty cool, and the application, Briar[0], is pretty neat.

I just installed it, and it wants to do a physical key exchange with a contact.

Time for a keysigning party!

[0] https://briarproject.org/

I wanted to try Briar but discovered that you have to be within bluetooth range of anyone you want to add as a contact.

I was planning to use it with my friend in another country but there's no way that can happen.

Please mark in the headline that this article is from 2016.

Isn't control over the application the new centralization?

Reply on news.ycombinator.com