Matt's Roof Garden

Powered by 🌱Roam Garden

September 1st, 2021


They weren't all together bad people but I found myself reflecting their cynicism and aloofness


How we're using @RoamResearch to build collective learning in the @code423n4 DAO


Networking by constellating

Differentiating between conversations and high-level thinking they have found to be key

They use a modified evergreens and seedlings framework from Person/Maggie Appleton

Evergreens: High level thoughts they are trying to figure out, container/parent thoughts

How can our website be improved? What is our reporting process? Things to know when onboarding etc.

Seedlings: individual ideas that make up the greater evergreens

Conversations constellating

They use Roam not only as a place for though but as a place to share how they do things like how they create out loud, how they create a welcoming community

They title pages in the Person/Andy Matuschak style, base the title on a belief that is had. One statement that summarizes a whole thought

Pages are titled based on how they will be useful in 6 months and how they will connect to other ideas to make up a larger evergreen

Tags are used sparingly and specifically

The tag is about the function of the idea and not what type of content it is

Tags are impossible because of bloat, the real value is in connections to Related Ideas ie seedlings

Always create a page on the DNP. This not only links to a date but also, since titles are statements of belief it allows for discussion and refutation

They try to immediately link a Seedling to an Evergreen

Pages are for metadata and most writing is in the linked refs. This is perfect if your ideas are atomic enough and avoids pages that quickly become outdated

Made some incredibly good progress on an API for the Annotation Department. This will enable moon+ to send highlights directly to the backend. Only problem I haven't solved yet is sequencing. App/Moon+ Reader sends highlights immediately after creation which makes concat actions difficult. I assume Readwise does it by collecting highlights before processing them but I have no clue how to implement that

The highlights also don't have any kind of sequence to them, only chapter title which will be a problem down the line