Cell Store Vocabulary
In this section, we drop all the vocabulary of the cell store paradigm (most of it is borrowed from the XBRL world).
Facts
A cell store contains a very big pool of facts, millions to hundreds of millions and potentially even more in the future. A fact is an atom of data. Each facts is made of a context associated with a value.
For example, a fact might report the value 231,839,000,000, associated with a context that specifies that it's the total assets owned by Apple on September 27, 2014, in US Dollars.
The context of a fact, also called dimensional coordinates of this fact, is made of pairs associating aspects to some values. In this case:
Entities and archives
Facts are reported by entities (such as Apple). Usually, entities report facts about themselves, so that it's the same entity that appears in the contexts of the facts that it reports.
Facts are not reported one by one, but are typically submitted to an authority like the SEC in archives. A bit like you don't buy pop corn one chunk at a time, but in a pop corn package. An example of archive is the fiscal report filed by Apple for the fiscal year (FY) 2014.
Components
Within an archive, facts are grouped in components. Components are facts that make sense together and are organized in data cubes. A fiscal report may contain, for example, a balance sheet, an income statement, etc.
A component can be displayed in a fancy way that resembles a spreadsheet or a pivot table. On the spreadsheets, some cells contain fact values (data), while others contain headers (metadata). The latter are called report elements.
There are six kinds of report elements: concepts, abstracts, line items, dimensions, members, hypercubes.
The report elements contained in a component are also sometimes referred as taxonomy, and are organized in a hierarchy called a model structure.
With this model structure, it is possible to retrieve all facts that belong to a component.
Another way of displaying a component is under the form of a fact table, which is a flat display of a data cube. However, this is less intuitive for business users and mostly used by developers to design pivot tables.
Besides its model structure, a component will also contain labels for each report element and in various languages. These labels are used those that the business user sees.
Finally, components contain rules that automatically validate facts, or automatically add new facts.