A protocol to power Antype's future? Or at least one to interop with

I just got an email about this from another PKM app’s mailing list:

Noosphere is a protocol for thought. A worldwide knowledge graph on top of IPFS.

  • Worldwide knowledge graph . Noosphere is a massively-multiplayer shared graph made up of smaller public and private graphs.
  • User-owned data . Apps trap thoughts in SaaS silos. Noosphere fixes this. Everything is syndicated to IPFS, so you have credible exit through any IPFS node.
  • Self-sovereign identity . Authentication happens through user-owned keys not ids, using UCAN.
  • Local-first sync . Everything works seamlessly offline. Files can be synced live to your device.
  • Decentralized over IPFS . Noosphere is peer-to-peer, making it a durable repository for human knowledge. Lots of Copies Keep Stuff Safe.

Noosphere is a permissionless protocol, like HTTP or IMAP. You can build whatever you like on top of it. We’re going to use Noosphere to power Subconscious’ multiplayer and sync features. We hope other apps find it useful too.

This sounds to me like a more fully (or at least publicly) articulated vision and functional protocol for doing what Anytype is aimed at doing. We already have Anytype the app, and it’s very far along in PKM functionality, but the protocol behind it, the Web3 nature of it, etc. is all still theoretical AFAIK. We currently sync to a centralized server. There are plans to decentralize, self-host, all of that. But it’s all in the future. Addressing other people’s content, etc. has not been clearly articulated by the team, certainly not to this specific a degree.

While Noosphere appears to also make some compromises (they call the servers “Web 2.5”, it’s worth reading details of their implementation), it’s notable to me that this already exists and is already open source. I hope the Anytype team is aware of this and is strongly considering how it interacts with their future plans! I would love to see compatibility in some way with this, if not simply to see Anytype’s infrastructure team jumping onboard this base protocol effort and contributing to their open source project. I can’t speak to the quality, etc. as a non-dev, so I totally get if it’s not up to par of what the Anyteam is building, but I do hope there is still room for interop. Competing standards can have some value, but they also create fragmentation, which is hard on users.

As long as all protocols, including Anytype’s, have “credible exit” as a core principle, we as users are in good hands though. :slight_smile:

7 Likes

Hello @Oshyan !

Thank you for pointing it out. You are absolutely right, it’s very close in the narrative to what anytype is building.

The protocol for thought sounds even more ambitious than a decentralized web OS. However, Noosphere is very early in development and we cannot find any ideas we can use from it. At the current state their protocol looks like the anytype protocol in the first version of anytype, three years ago. Just a sketch.

I believe it’s not a fair statement that the anytype protocol is theoretical. Noosphere is much more theoretical than anytype is.

The anytype protocol exists and powers the anytype app. Our protocol supports e2e encryption on top of a distributed P2P database (Noosphere does not even have a theoretical answer to this problem yet - see theQ&A part of this) video). Anytype protocol today already supports thousands of people using anytype clients every day. The code is just not open yet, and neither is the code of their app.

We intentionally prioritized practical development over publishing of conceptual papers, as we believe that we should only publish these once we have demonstrated that this architecture can be achieved in practice.

The anytype protocol and app are decentralized. The way we run anytype sync-nodes right now is centralized, because it is in alpha. After you install it, you can use it without the caffe-node. Later, you will be able to run your own.

What Noosphere proposes is to launch a network of classical servers on top of IPFS, which will read-write IPFS and serve this content to clients via API. They call it decentralized, but it is still web 2 with backups on IPFS. Too many questions here.

From the product perspective their current design is made only for public publishing of static data, not for private db-powered usage. It would not work for messaging and other high performance p2p applications. Anytype on the contrary, can do all of that and even more. Better data model, markup, and as you fairly mentioned - the app.

What I believe is missing is our presence in the public space. We are going to fix this asap and start publishing articles about our philosophy and design decisions. And of course, we are going to open the code. So please be sure that “credible exit” (aka data portability) is our principle.

8 Likes

Thanks Anton for that in-depth and thorough reply! Much appreciated. I agree that the openness is the main positive thing that Anytype is lacking in to-date, while it has a much more developed client/app, and as you say, more developed back-end/protocol/etc. It only remains to show everyone how it all works and get people in to start building on it themselves. Can’t wait to see that start to happen. :slight_smile:

3 Likes

Hi @Oshyan @anton
this is an interesting thread and I’m actually happy to see another high profile project innovating around similar concepts and ideas as that is a very strong confirmation for anytype’s value propositions.

Behind this thread lies a deeper question in my opinion which will increase in relevance with more related projects joining the space.

What are the trade offs of opening up anytype vs. developing it with a focused team in a rather closed fashion? This refers to open sourcing the code base but also to being more of a community centric ecosystem with more public visibility (developer conferences, hackathons etc.). The answer to this question will very much define the DNA of anytype going forward, it is quite foundational.

There are various examples of successful projects and companies sitting on either end of the spectrum - no silver bullets or easy answers. The obvious advantages are

Closed: + moving fast + efficiency + product velocity + focus + reduced complexity + harder to copy
Open: + developer mindshare, buy-in + diversity in perspective + virality, fly wheel effects, entry barriers through 1st mover advantage

Personally, I feel that a gradual approach might make sense. Why not consider to open source some components of the stack where we feel that external developers could contribute or provide new perspectives? Why not running an anytype conference format for developers (IRL + virtual) to spark + maintain community engagement?

Curious to see how this thread and the topics touched on will evolve over time.

See you soon!

3 Likes

Parts of the app will be open-sourced with the public launch in about 5 months.

1 Like