The Read-Write Web - A Common Purpose

I just read this lovely article about the future of the Web by @PaulFrazee creator of Beaker Browser and the Dat project.

Several threads weave together here to highlight the common underlying themes of different projects, all of which owe much to the insights of the web’s creator, Sir Tim Berners-Lee (TimBL). His yet to be realised vision of a read-write web, was I think predicated on the assumptions of ‘universality’ which Paul’s article explains with admirable clarity.

Since becoming a Maidsafe / SAFEnetwork fan I’ve been exposed again and again, but in different ways, to this fundamental idea which was TimBL’s innovation, but which was not properly taken up by those not always building on the standards he strove to develop. Paul highlights examples of this ‘sabotage’, including Adobe Flash and reddit upvotes. He presents a test for ‘universality’ which asks Would this feature function or have meaning if it were disconnected from a specific service?

If the answer is yes, he regards the feature as universal and suggests CouchApps as the nearest we’ve come so far, to this unrealised goal of the web’s inventor. Ding! CouchApps were the first thing I studied when trying to find a model that would make it quick and easy to build on the universal platform that is SAFEnetwork :slight_smile:

Paul mentions how authentication was one blocker to the further development and adoption of CouchApps, and indeed I can see that solving this is a key element of all the other efforts to create a universal read-write web that I’ve looked at so far:, Project Solid (TimBL) and of course SAFEnetwork.

I’m not familiar with how Dat solves the authentication issues, but Paul mentions that it does.

In each of these projects there’s a separation between application and storage, a standard API for storing and publishing, decentralised ownership and control of data by the user rather than a third party, independence from service providers (no walled gardens or lock-in), and an element of data standardisation.

Each project pushes the envelope in one or more of these areas, and in each case I think there’s an opportunity for all projects to both learn from the others to strengthen themselves in one area or another, and for them all to work together to find common ground (using or supporting common standards for example) to amplify the capabilities and impacts of each other.

For example, TimBL’s Project Solid excels in defining and applying Web standards for authentication, storage, schemas, and so on, and being most compatible with existing Web infrastructure.

I don’t know Dat well enough to be sure of it’s greatest strengths, but it occurs to me is that it lies somewhere between Solid’s compatibility with existing infrastructure, and SAFE network’s maximal “from the ground up” solution. My impression is that Dat is relatively easy to deploy compared to either of the others.

In its turn, SAFEnetwork maximises security and longevity of data, privacy, and universality through a truly autonomous, decentralised data storage and communications network.

I’m hopeful that those involved in, or with greater knowledge of the other projects - not just those I’ve named, but any that share the goals of a universal read-write Web - will be interested in commenting here: perhaps to suggest what they regard as the strengths of a particular project, or any areas where it can learn from the others, and where projects could adopt compatible approaches which leverage each other.

In the last instance, I give the example of SAFEnetwork looking to support the LinkedData Platform (LDP) API through a JavaScript compatibility API, making it easier to write apps which work with Solid or SAFEnetwork.

In time I think we can go much further than this, but having reached this vantage point I’m keen to demonstrate some basic capabilities before getting ahead of myself.

At this point I am very optimistic for the realisation of TimBL’s vision in one forum or several, and very excited to be here, doing my small part, at a time when this is much needed and seems so close.


Another good piece in a similar vein to the Ruben Verborgh article you posted earlier. Three Shifts To Decentralise The Web . Is it your understanding that the authentication issues mentioned are, in the example given, that anyone can come along and delete my ‘likes’ because there is no central server-based authentication service to deny them that ability?


I don’t know what was the issue with CouchApps and authentication, but I suspect it was needing per app/service/server credentials rather than a portable user ID across services.

1 Like

UPDATE: I have an early (at this time incomplete, not well tested but working) implementation of LDP in the browser as part of my SafeNetwork Web API (nodejs in the browser).


Above is master branch, but see branch ‘wip’ because I have not yet pushed the working LDP to master yet: