Build a collection that answers a question

A collection becomes valuable when its selection and order help someone accomplish something. Before adding another item, write a one-sentence purpose: understand a controversy, teach a concept, compare versions, or prepare a source-based explanation. Use that purpose as the collection's admission rule. 1. Name the outcome. A title such as early-web interfaces is a topic; a description such as compare how three interfaces invited participation tells a reader what the collection does. Keep the description visible in your notes even if the software offers only a short name field. 2. Give records distinct roles. Choose an entry point, essential evidence, a contrasting example, a source that complicates the story, and an optional deeper reference. A useful small collection may need fewer items than a large unstructured one. Remove near-duplicates unless the differences themselves are part of the question. 3. Use tags for reusable facets. Topics, media types, and methods can cross collection boundaries. Keep tag spelling consistent and avoid creating several near-synonyms unless they mean different things. A record can serve more than one collection without being copied into several independent versions. 4. Order for the reader. Start with enough context to understand later material. Put a source before the claim that depends on it, or explain why a surprising example appears first. If the software does not support a custom order, keep a numbered reading note with record identifiers. 5. Add an annotation for each essential item. Explain what the reader should notice and why it belongs. A sentence such as this launch announcement establishes the feature's public description is more useful than important source. Include any limitation that could cause the source to be misread. 6. End with a task or open question. Ask the reader to compare two examples, inspect a source, or explain what evidence would change the interpretation. This turns passive browsing into a small piece of learning or research without inventing a quiz score as a measure of understanding. Review the collection after a week. Can you still explain why each item is there? Are all essential links usable? Does the selection omit a relevant perspective or depend entirely on one source chain? Keep an overflow list for interesting material that does not belong in the current route. The finished collection should be exportable as a short description, ordered identifiers, annotations, and sources, so its intellectual structure survives a change of application.

Editorial review: 2026-09-23. New evidence may change this account.

Report a problem with this record