5 Information Management
5.1 Research work produces more than data
A research lab also produces:
- questions
- reading notes
- decisions
- protocols
- to-do lists
- dead ends
- interpretations
- publication drafts
If these are scattered across email, chat, browser tabs, and random files, the lab becomes hard to trust.
5.2 Build a layered information system
Use separate but connected layers:
capture- a low-friction inbox for quick notes, links, ideas, and tasks
reference- curated notes, reading summaries, protocols, and key concepts
project context- notes tied to a specific project, dataset, or analysis
decision record- short records of why important choices were made
publication- polished outputs for sharing with others
5.3 Prefer durable formats
Whenever possible, keep important material in formats that are:
- plain text or open standards
- easy to search
- easy to export
- resilient outside one vendor platform
Markdown, CSV, YAML, JSON, and PDF all have limits, but they are durable enough to support long-term work.
5.4 Link notes to artifacts
The more technical the lab becomes, the more important it is to connect narrative context to computational artifacts. Notes should point to:
- repository locations
- dataset versions
- scripts or notebooks
- figures and tables
- external references
- open questions
Without these links, your information system becomes a diary instead of a lab notebook.
5.5 Create a minimum note taxonomy
Keep this light. A helpful starting taxonomy might distinguish:
- fleeting notes
- source notes
- project notes
- methods notes
- decision notes
- publication notes
The goal is not a perfect knowledge graph. The goal is retrieval.