Synereo Weekly Update

HyperSpace

Published in

Synereo

8 min readFeb 7, 2016

Update Summary

This week, we talked about the new Taiga task management system starting to be employed by the Synereo community.

Updates were given about the work on Casper, the new Proof of Stake mechanism, Synereo’s back-end development, the Synereo MVP, and many more.

Contents in Brief

00:54 — Individual updates

02:59 — Ed, LivelyGig
05:08 — Synereo MVP
12:15 — Taiga management system for Synereo
13:33 — Cohorts — MaidSafe, Safe Exchange, Ethereum
15:54 — Scalable Blockchain: Proof of Work, Sharding, Special-K
18:45 — Scala Apprenticeships
24:03 — Casper and Shards
29:25 — Proof of Stake
31:30 — Investor
32:18 — More maths: Type systems
41:00 — Active & Passive browsing, Attention Economy
53:08 — More on UI and MVP
57:26 — Crowdsale/BnkToTheFuture

Detailed Notes

Jed communicated that he has been staying out of the hangouts to leave room open for newcomers. He’s been active behind the scenes, answering questions via chat and coordinating the youtube presentation. Currently, he is establishing the role of Synereo Ambassadors, a new opportunity for others to earn AMPs for helping the community grow. Dor hinted at developments with the weWOWwe network, and will fill us in more next week.

02:59 — Ed, LivelyGig

Ed talked about the North American Bitcoin Conference in Miami last week, and some inroads were made regarding LivelyGig, and Scala developers. He and Ryan are also working on the UI, and an end-to-end demo. Nirvanic is working on wiring up the diode and flux architecture. The first target market for blockchain and smart contract development has been chosen, and they are moving forward on that.

05:08 — Synereo MVP

One of Ed’s questions was about the initial target market for Synereo. Christian suggested checking into iTunes podcasters as a potential group to review. Eric said that from the UX perspective, target market is the most important question. He & Dina think the blockchain and “hacktivist” communities are good initial targets. They understand the technology, and can be easily led to finding uses for Synereo. He & Dina spoke with Noy about design, and the user’s story, leading to an MVP that will be a great demo of Synereo’s capabilities.

12:15 — Taiga management system for Synereo

HJ said that he had found 2 Scala developers, and they are thinking of having a workshop on Synereo. He also started a page on the Taiga project management site, to see how that might help our project. Synereo on Taiga

13:33 — Cohorts — MaidSafe, Safe Exchange, Ethereum

Rich says he loves what Synereo is doing, and follows & promotes it, and other similar projects like Maidsafe and Safe Exchange. He thinks they will all grow faster if we all work together. Greg agreed, adding that there is an abundance of work for all.

Rich also asked about Casper and Ethereum, and how Synereo’s work with them will benefit Synereo. Greg explained that the existing stack that Synereo had was largely a content delivery network, and that they were hoping that better blockchain technology would be produced somewhere (that would address the limitations of Bitcoin’s chain).

15:54 — Scalable Blockchain: Proof of Work, Sharding, Special-K

Bitcoin’s current algorithm, called Proof-of-work, did not appear to be what was needed. This is what led Greg to first work with the Casper group in order to meet the commitments that Synereo had with its investors. Greg reports that Casper plus Sharding should make the blockchain scalable. Adding Special-K into this tech-stack, we’ll have a metered, monetized, content delivery network.

18:45 — Scala Apprenticeships

Greg met with Gary Stevenson, an experienced functional programming dev in Sydney, about being an apprentice working on the backend of Syenero in order to learn Scala. Gary’s also very interested in freelancing through LivelyGig. Greg and Gary have been working on improving the performance of the Synereo back-end.

Greg went into some detail on the code for the LivelyGig UI.
When you deploy a node the UI-code and the Server-code are on the server. That should be the case if you have support for course. If you can separate the two, then it makes deployment easier. Also the LivelyGig importer has been improved, by upgrading Spray to a newer version. So the GlosEval code and their dependencies are now ported to the newer version of Spray. The SBT-built is now also more modular and doesn’t get mixed up with the dependencies.

24:03 — Casper and Shards

Sharding divides the data set and distributes the data over multiple servers, or shards. Each shard is an independent database, and collectively, the shards make up a single logical database. See about PoS Vlad Zamfir’s or Vitalik Buterin

It seems to be better to think about Casper and computational models that support Sharding at the same time. You can shard your database in different ways, for example according to: the application state, the geographic location or the name space (the address state).

What Greg keeps telling them is that the namespace or your address space is the most natural sharding.

29:25 — Proof of Stake view of the blockchain, and how it can represented as a traced monoidal category.

Another thing is the betting cycle and the whole Casper view of the blockchain. This can be nicely represented as a traced monoidal category This is a category with some extra structure which gives a reasonable notion of feedback.

You can look at the convergence of the betting cycle as an iterated function system. Convergence corresponds to having attractors for the iterator function system. And then you can prove things about convergence if you set up the economics of punishments and rewards appropriately. So if you have an attractor in line then you can construct an iterated function system that has that attractor as its attractor or as its fixed point.

So you can look at the bets that come in and see how well they match against the iterated function system that has that attractor. And if they stay too far from the behavior that is dictated by that iterated function system, then you punish them. That gives you a maximum of freedom. You’re free how to place your bets as you want. But bets that don’t line out with something that at a minimum converge to a winning block at a particular type, are gonna get punished. So the only validators that are left standing, are the ones that conform in some way to something that’s going to converge to picking out blocks at a particular height. This a rough sketch of how to make the Casper betting cycle converge.

Then the other observation is that the betting cycle itself is a trace in a mathematical structure, called a trace monoidal category. See a paper by Masahito Hasegawa which make that argument crystal clear and precise. There’s a function that turns one into the the other and the other into the one.

That combined with the notion of linear types that gives you a nice way to decompose the blockchain in a scalable fashion. And there Greg present a contracting language that already has sharding semantics in it. And then the type system supports the sharding semantics. And that allows you for example to control computational power, because well typed contracts always terminate.

Linear logic provides a kind of natural programming model for reasoning about sharding the blockchain. And from Casper it’s argued that the blockchain can be viewed as a traced monoidal category. And a traced monoidal category are a natural place to interpret computations based on linear logic with its connection to the pi-calculus. And Greg has found a way to interpret the betting cycle in terms of the pi-calculus protocol. So all the pieces are fitting together. So we’re slowly piecing together a coherent view of how the blockchain is constructed out of the mathematical building blocks.

31:30 — Investor

They met with a potential investor who is supportive of the work merging Special-K and Ethereum.

32:18 — More maths: Type systems

The relationship between Type systems and Computational models that Greg has been working on for the past 7 years, and the algorithmic mechanism that takes data and generates type systems, is making progress. He and Mike Stay have been writing up results in higher Category Theory, in order to explain how this ties into other, well-known results to that community. This was a woo-hoo moment!

Finally, he shared a site that he really loves, called Hit Record, from Joseph Gordon-Levitt.

41:00 — Active & Passive browsing, Attention Economy

The conversation then touched on the difference between ‘passive’ and ‘active’ browsing experiences, and how Synereo will address both, adding more collaboration to what existing networks offer. Christian added that this addition really shifts the paradigm (away from the centralized ones of modern society). Greg topped that off by saying that this is the idea of our Attention Economy. If we start with an environment where devs can hack the platform with the platform, the platform will expand in many ways. Eric talked more about the attention economy, and how it works for people, in regards to these kinds of network projects. Artists and other content creators will come onboard organically, after the devs build creative spaces and tools for them.

53:08 — More on UI and MVP

Prompted by a question from HJ, Eric covered the ‘storylines’ idea for users again, and starting the MVP (mentioned at the top). Dina will be continuing where she left off from her presentation last week soon.

Ian has been playing around with and trying to learn Scala JS, because he wants to get started hacking stuff on Synereo and LivelyGig.

57:26 — Crowdsale/BnkToTheFuture

Christian asked about the next crowdsale, which Greg answered by saying that the BnkToTheFuture project will be starting very soon, and there will be an interview with them soon (it was scheduled, but had to be delayed).

Scala developers are most welcome and will be rewarded in AMPs!

Please read the Open questions for pitch deck

Some links from the chatbox:
Access control
Sessions

That’s all for this week!