Multidimensional+Model

=Multidimenstional Model= The multidimensional model is a logical model leverged by digital financial reporting. (RDF/OWL Ontology)

The multidimensional model is an approch to modeling information. The relational model is another model. Each has its strengths and weaknesses. The following give you information which will help you understand the multidimensional model and how the US GAAP Taxonomy, SEC XBRL Financial Filings, and other types of digital financial reporting use the multdimensional model.
 * A **dimension** is simply a **characteristic**. Condider this number: 1000. Given only that number, what can you tell someone about that number? Not much. But if you provide more characteristics of that number it has more meaning. For example, the value of cash and cash equivalents expressed in US Dollars as of December 31, 2010 for the legal entity, a consolidated entity, ABC Company which is the SEC filer with the CIK of 000000001 is 1000, the items in red being characteristics of the value, provides more information.
 * A **scaler** is a value which has no other characteristics which are necessary to understand the value (i.e. it needs no dimensions). For example, the value for pi is always 3.14. You need no dimensions to understand pi.
 * A **table** has two dimensions; rows and columns. Another name for a table is an array or a matrix, or more accuratly a two dimensional array or a two dimensional matrix.
 * A **cube** is a three dimensional table, array or matrix.
 * A **pivot table** is another name for a cube, table, array, matrix.
 * A **hypercube** can have any number of dimensions, or "//n"// dimensional table, cube, array, or matrix.
 * A [Table] in an SEC XBRL financial filing is really a hypercube which could have any number of dimensions or [Axis]. The US GAAP Taxonomy and SEC XBRL financial filings uses the term [Axis] to refer to dimensions. **Hypercube** is what the XBRL technical syntax calls what the US GAAP Taxonomy calls a **[**Table**]**. Basically, a hypercube is an "n" dimensional table, or a table with an infinite number of dimensions or what the US GAAP Taxonony calls [Axis].

To better understand this information above, the additional information below is helpful.
 * **Everything modeled in an SEC XBRL financial filing is modeled as a hypercube, either explicilty as a [Table] or implicitly.** That may seem odd, but it is true. Every fact value reported has at least two axis: Reporting Entity [Axis] and Period [Axis]. Why is this? Because they are required by both XBRL and by the SEC. The XBRL context element has an element called . This articulates the CIK number of the SEC filer, the "reporting entity". The XBRL context element articulates the calendar period of the fact value. These are not called [Axis] or dimensions, but logically they act exactly like any other dimensions/axis.
 * When you **render information** from an SEC XBRL financial filing, sometimes you might want to see certain information in columns as opposed to rows; other times you would prefer to see information in rows rather than in columns. How you render or present the information does not change the information. Defining information based on how you want it presented is fundmantally incorrect. XBRL is called "interactive data" by the SEC for a reason; you can easily change the presentation. This is similar to how a pivot table works.
 * There are even **implied** [Axis]. For example, if you don't provide a Legal Entity [Axis], it does not mean that the entity has no legal entity; clearly it has a legal entity. If it is not explicitly provided, in many cases (most or all cases) the legal entity is implied to be the consolidated entity.
 * It is better to be **explicit**. Your legal entity might be the a parent holding company but if you don't explicitly state that, someone may misinterpret that you mean the consolidated group. Computers are not very good at implying information, it is best and safer to be explicity whenever possible.
 * Business Intelligence (BI) systems use the multidimensional model, however:
 * There is no globally standard definition of the multidimensional model (i.e. each BI vendor implements this model slightly differently. As such, BI systems are not generally interoperable.
 * BI generally "forces" the OLAP model (online analytical processing). Sometimes you need OLAP, sometimes you don't.
 * BI is read only, it is not read/write.
 * BI does not handle textual type information well (text, narratives); it works best with only numbers.