The objective of XBRL is the semantic representation of financial statements. Rendering is the objective NOT of xbrl, but of a display language such as XSL/XSLT or a page description language (such as pdf). The best example I can give is that of IRS, probably the earliest and most effective implementor of SGML. IRS stores all forms just once, but renders them in a whole bunch of formats as necessary.
Rendering denotes FORM, not substance of the document. It is the xbrl (semantic tagging) tagged document that denotes substance of financial reports. I don't know what the AICPA guys were smoking when they wrote that.
A document that is filed with the SEC is THE legal document, and derives its meaning from its interpretation under law. That being the case, mucking around with rendering is a red herring we could do without.
The footnote tags in xbrl have not found acceptance. That is so because of the reasons I have given.
We need to look at other applications where there was success. I gave two examples: UPC codes and EDI.
Until we develop lexical resources to help practice as well as filers, we have to live with this heap of needless and ambiguous information.
I have always thought that in academic research those who use massaged data (that includes compustat, audit analytics, WRDS interface and so on) are either lazy or not-yet-really-ready-for archival-research, especially when there is primary data available for free. As department chair I fought like hell to prevent our scarce department resources used to buy stuff like WRDS and support faculty laziness (the accounticians went past me to the Dean and got the money from the school funds; I refused to have them taken out of department funds).
There is wealth of data in the EDGAR filings that very few of us have bothered to look at. If our faculty can not (and do not) have the competence to examine such data after four or more years of doctoral education, I do not have much faith in such education.
No comments:
Post a Comment