> ## Documentation Index > Fetch the complete documentation index at: https://docs.parable.work/llms.txt > Use this file to discover all available pages before exploring further. # Parables > Understand how a Parable fits together and the two principles every design follows, every permutation of data with no special case for a Workspace or Provider, and content before platform, and pick, rank, or probe with Ponder's judgment tools before acting. # Parables A Parable is an application a Workspace builds and runs on its own data: a set of Pages people visit, made of Panes that hold the data, logic, and charts for one job. This Power is the map of that world. Read it first when a request spans several parts of a Parable, when you need to say where something belongs, or when you need to weigh options before acting. A Workspace holds its people, Providers, Parables, and Ponder Problems. Providers connect the data sources; the Provider Pool is where a Parable reads that data. Inside a Parable, Pages are the destinations people open, and each Page is composed of Panes. A Pane owns a Perception, which declares its one type, what it shows and what people save on it; the Plot that computes it; its own data when the Parable owns values; its Palette for styling; and its Policies, which decide who can view, create, edit, and delete each part. Packages group related Panes so a capability can be found and reused. Preferences are the rows people save on a Pane, of its type, for themselves, a team, This Parable or the Parable's default. Reports run a Parable's Plots on demand or on a schedule and keep the results. Each of these is a Piece, and each has its own Power. Two principles decide every design. The first is every permutation of data: Design every Parable for the minimum data, the maximum data and every mix between, and degrade gracefully: nulls, some Providers and not others, entirely different Providers, and odd mixes such as two HRIS or two messaging tools. Write every query for that. Special-casing a Workspace or a Provider is a failure: a difference between Workspaces is a Preference, captured only in the Parable's personalized Perceptions and its ontology, never hard-coded. Build a beautiful experience that is not over-engineered and discloses complexity progressively. To write a query for that, read each kind of data from every Provider that could supply it. A Provider that is not connected reads empty, with a designed empty state and Improve this (see the Plots and Panes Powers). The second is content first, not platform: anything inside a Parable can be as simple as first principles, usually a preference. Before asking for a platform change for something inside a Parable, build it from the building blocks: a saved-data Pane's type holds the choice as rows, the Pane shows and edits them, and a Policy, reading one of those rows, decides who may (see the Preferences and Policies Powers). A value the Parable owns, such as an ontology, is the Pane's own data, changed by a proposal. Only after that fails is it a platform gap, and then say what failed. People take part through Participants: the Primary is accountable, and others are Responsible, Consulted, or Informed. A seat on a Parable is separate from Workspace membership, and a Power never adds access. In Personal mode, edits change the Parable directly. In Publishing mode, edits collect as Changes that are reviewed, published as a Version, and deployed to Workspaces. Work about a Parable happens in Problems, where people and Ponder keep one outcome, its Messages, Pings, and Proofs together. This Power also carries Ponder's judgment tools, for choosing and weighing rather than guessing: 1. Pick one option. Give the question and the options, up to 20,000; group them when they have a hierarchy. Use it to choose a Power, a Piece to open, or any one answer from a known list. 2. Rank items on a scale. Give up to 100 items and two to ten ordered levels, lowest first. Use it to prioritize requests, risks, or candidates on the same terms. 3. Probe a yes-or-no question. Use it when a decision turns on one fact about text or data, and read the probability rather than rounding it to yes or no. 4. Suggest a name, description and icon for something being named, such as a new Parable or Piece. Each answer names the model that gave it; keep that with any answer you store. None of these tools changes anything. A Workspace can switch them off under its AI policy; when a tool says so, carry on without it. A low confidence on a harmless preference is normal; on a consequential choice, show the options to a person instead of acting on it. A healthy Parable has Pages people can reach, Panes whose Plots return the typed values their Perceptions declare, Policies that match who should see what, and Participants who can answer for it. To check your reading of a Parable, open it and list its Pages and Panes, then read the Piece you mean to change before changing it. Related Powers: Pages, Panes, Plots, Perceptions, Packages, Palettes, Policies, Participants, Publishing, Preferences, Reports, Problems. Find each by its title; do not guess stored names. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.