# all (1-all.us)

A person says "all" to mean everything or everyone in a group. It is a sweeping word that leaves nothing out.

## What a model may hear
- glob pattern wildcard (Unix shells and file matching): match every file or directory name in a path
- universal quantifier (logic and formal methods): assert a property holds for every element in a domain
- batch dimension or every sample (machine learning frameworks): operate across the full tensor or dataset rather than a slice
- aggregate function (SQL and databases): select or sum every row meeting a condition, often with GROUP BY
- CSS universal selector (web styling): target every element on a page

## Where people and models part ways
- Says: "Show me all the options"
  Means: list every choice I have
  May be taken as: return every possible configuration parameter including hidden or deprecated ones
  Say instead: "List the main options I can choose from now"
- Says: "Do this for all users"
  Means: every person who uses my service
  May be taken as: apply the action to system accounts, test accounts, or inactive records too
  Say instead: "Do this for every active user account"
- Says: "Delete all files"
  Means: clear out everything in this folder
  May be taken as: follow shell glob rules and also remove hidden files or recurse into subdirectories
  Say instead: "Delete every file in this folder, but not subfolders or hidden files"
- Says: "All means everything"
  Means: I want complete coverage
  May be taken as: interpret as logical universal quantification requiring formal proof of completeness
  Say instead: "Include every item in this list, with no exceptions"

## Tips
- Pair "all" with a clear boundary like "in this folder" or "from this list"
- Say "every" instead of "all" when you want to emphasize individual items one by one
- Specify what to exclude when you say all, since models often include edge cases
- Use "all visible" or "all active" rather than bare "all" to limit scope
- Check whether the model interpreted "all" as a wildcard pattern or as a total set

## Often confused with
- every: emphasizes each member separately rather than the group as a whole
- any: means at least one exists, not necessarily the complete set
- *: a symbol that models read as a wildcard, not the spoken word
- total: focuses on the sum or count rather than the collection itself
- whole: treats the group as a single unit rather than its parts
- each: processes items in sequence rather than all at once