The world wide web is great to publish and find information, but it suffers from fake news and quality journalists not getting paid for their content. Wouldn’t it be great to finally have an editor that allows authors to charge micropayments and quote their sources?
This post is inspired by @vrandecic sharing “Project Xanadu – The Internet That Might Have Been” and seeing @nichtich picking it up.
Right, Denny, and it should be connected even more.
Technically, the web is a perfect basis to share information and it can well be extended to add the parts from Xanadu that were identified as making it more humane:
- micropayments, for example using W3C micropayment initiative
- backlinks (transclusion) using something like the ActivityPub shares collection but for longer texts.
I already wrote that I am willing to pay for content, and the more people pay for content they want to see, the more positive content will follow. I also want a Semantic Web editor. In this post I share some thoughts about this.
In a post-factual world where assholes (as defined in “Team America”) use lies, deceipt and manipulation to start wars, win elections, deny climate change, we need to nourish a culture of linking to sources and spreading facts.
We can do this by using RDF markup.
The RDF/Semantic Web/LinkedData enables us to add metadata about sources, backlinks, shares to content. We should add the “source” to our copy-pasting operations on a metadata, operating system level. When we keep the context where a statement was copied from (in a quadstore or named graph), we keep the link to the source. Storing it on a distributed semantic web (evolving the distributed Fediverse architecture) we can have global backlinks on a shoestring budget.
We need end-user editors that enable people to do this … and browsers that let people query/filter/remix the data … and query more data from the sources.
Currently, the Semantic Web is in full speed and quickly adopted as marketers doing Search Engine Optimization (SEO) have an incentive to write schema.org markup. SEO people often only care about Google and LLM readers, not about end users.
Reading The Internet That Might Have Been, it reminded me of the time when Stefan Decker and me started “Semantic Desktop” in 2004, it felt the same “cult”. Stefan convinced a “Doug Engelbart (Bootstrap Institute, USA)” into joining our program committee for the “1st Workshop on The Semantic Desktop – Next Generation Personal Information Management and Collaboration Infrastructure“. This workshop on the 6th November 2005 and the many others that followed had a similar vibe: based on open source and standards we wanted to connect all knowledge into a Memex / Xanadu.
I realized I live by what Christiana Figueres, a diplomat who negotiated the Paris Climate agreement, describes as “stubborn optimism“: Optimism means envisioning our desired future and actively pulling it closer. I was actively pulling Memex and Xanadu closer into reality.
Plom once said to me that he started Plomwiki in a day and it worked on day one. He learned that simple systems, that work from day one, survive. Xanadu was too complex and ambitious and the “it will ship in 6 months” project management didn’t help. Chandler had the same fate, documented in Dreaming in Code.
Semantic Desktop was just right when it came to complexity, we built and shipped it in 3 years spenging 16 mio EUR. It reached about 4.5 million users and then collapsed, partly due to some bad decisions I made. NEPOMUK was and end-user application with an editor and database for the semantic web. In the hands of users on desktop and mobile systems, it allowed everyone to benefit from the metadata and understand the value. Being part of the operating system, it was a good foundation for Memex-like features.
Back when the Semantic Web started, there was Closed World Machine, CWM, a commandline helper for a lot of day-to-day jobs. Inspired by Plom, I started hacking a Obsidian plugin for myself. I coined it “LinkedDataNotes”, it allows me to use linked data in personal note-taking. I want to link to people from my addtress book when I mention them in my private notes. I want to use semantic web data markup to enter the GPS position of diary entries. Backlinks and context are a nice feature to have at your fingertips. Currently it is 649 lines of TypeScript and it just plugs a quadstore into Obsidian. It falls more into the category of a Plomwiki than a Xanadu or a NEPOMUK. I wonder if anyone has the same itch to scratch. I don’t spend much time with this, as I volunteer for FridaysForFuture.at and Burners.at next to my job at portatour.com.
I would really appreciate an end-user editor for the semantic web to close this Memex/Xanadu-shaped gap in the world.