Re: XBRL object orientedness. multiple rollups.

[Chris Nelson <Chris@cnelson.demon.co.uk>]

I won't repeat the original message here, as it was very long.  It concerned
the XBRL standard put together by the accounting profession.  From the
description given in the e-mail, each XBRL "concept" in their taxonomy is a
different XML element.
I can assure you, with 12 years of experience of multi-dimensional data
reporting, this is NOT the way to go.  The international central banking
community (incuding European Central Bank, Bank for International
Settlements, EU members states Central Banks, Federal Reserves USA,
Australia, New Zealand, Bank of Canada, Japan, IMF etc.) and the European
Statistical Offices have been exchanging multi dimensional data and time
series data for years.  They have developed a model which has been re-worked
as a UML model: it  can handle any multi dimensional structure.  The essence
of the model is:
Metadata
1) Each multi-dimensional structure is described in terms of components or
"reporting concepts" (in a tabular form this might be represented as a
column heading).
2) Each concept of a structure can be linked to a code list or
classification - the real concepts.
Data
3) Reporting comprises just the multi-dimensional key and the numbers.  The
key comprises a ordered sequence of code values, one for each of the
components of the structure.
4) Additional attributes can be applied to any key - these attributes can be
applied to a rollup of the multi-dimensional structure.

The metadata only needs to exchanged when it changes: the data references
the metadata structure it uses.

This model will be input to the ebXML process very soon. I will be at the
Japan meeting and am willing to show the model to anyone who is interested.
Regards,
Chris Nelson
Dimension EDI