# Resources and datasets

> Learn how BisQue stores data, metadata, collections, and access rules.

## Resources

Resources are the durable objects you browse and share in BisQue. Common resource types
include:

* **Images**, including multidimensional scientific images;
* **Files** that BisQue stores without an image-specific interpretation;
* **Tables** with structured rows and columns;
* **Datasets**, which collect other resources;
* **Models and modules** used by analysis workflows; and
* **Workflows** that describe repeatable analysis.

Each resource has an identifier, owner, permission, creation metadata, and a hierarchical tag
tree. The API represents tag values as JSON scalars, so numbers and booleans do not have to be
flattened into text.

## Datasets

A dataset references canonical non-dataset resources in an explicit order. Membership is
direct and one level: a dataset cannot contain another dataset. Adding an item does not
duplicate its bytes, so the same resource can participate in more than one collection while
retaining a single identity.

Use a dataset when you need to:

* process a group of inputs together;
* preserve the order of a scientific series;
* provide one collection as a workflow input; or
* share a curated group without moving the underlying resources.

## Permissions

Choose **Private** for data that should remain visible only to authorized users. Choose
**Published** only when the resource is ready for the visibility rules of your deployment.
Project membership and resource permissions are related but distinct: access to a project does
not silently rewrite the underlying resource's ownership or publication state.

:::caution
Confirm your organization's data-handling policy before publishing sensitive or regulated
scientific data.
:::