You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We currently support embedded entities by way of headings or bullet points. It would be useful to be able to enumerate a set of solutions, risks, assumptions, decisions, etc. In the form of a table instead of being forced to use a bulleted list. This would involve extending the current markdown parsing to parse tables and extract from those tables a set of node entries with props based on the the column headings (eg title, status, ...).
We'll need to come up with a good way to determine the node type as that should be expressed once. we might need to try and detect a heading that's annotated with a type as we do already but then detect when there's only a table under that heading, signalling a list of entities of that type (in that case the heading content is otherwise ignored, it's not a title). Heading to know what type the table rows represent
Regarding tabular parsing - we could look in the first cell of the header row as a way of inferring node type, or base this on a natural language heading that follows the hierarchy level pattern (depending on implementation of #27). This would allow automatic type detection while maintaining flexibility.
Activity
mindsocket commentedon Mar 9, 2026
Regarding tabular parsing - we could look in the first cell of the header row as a way of inferring node type, or base this on a natural language heading that follows the hierarchy level pattern (depending on implementation of #27). This would allow automatic type detection while maintaining flexibility.