Digital Financial Reporting

The following are two links to information which represent further progress in creating a UML model for financial reporting:

• DRAFT of UML Model: This is a PDF with the images from a UML model created plus narrative added which explains the model. This model was made available by Cliff Binstock of XBRL Cloud.  Cliff modified the model based on suggestions by me.  The model is made available under a Creative Commons license, anyone can make use of this however they see fit, everyone is encouraged to contribute to creating and completing it.  However, it is best if one model can be agreed upon.  The model was created using the free ArgoUML Tool.  If you want a copy of the model which can load into that tool, please send me an email (CharlesHoffman@olywa.net) and I will send you a copy which you can edit.  Those who add value to this model will be included as creators of this UML model. 
• DRAFT Mind Map: This is a mind map of the same information.  What I am finding is that both of these serve a purpose.  The UML model is more concise and formal, but harder for business people to understand.  The mind map is easer to understand, but not "standard" or detailed enough for technical people (i.e. that is why UML was invented).  Therefore, both will be maintained.

The next step is to tune the existing draft and create additional "logical" and "physical" models.  All three of these models should fit together.  Personally, I also want to express the model in OWL.  The intension of these models is to help make XBRL for financial reporting unambiguous, provide semantic consistency, in financial reporting environments (domains) were XBRL's extensibility is utilized (filers of financial reports can modify the information model).

Here are some great examples of searching and using information.  Why can't searching for and using financial information be more like this:

• Finding a diamond: Nice interface for finding a diamond. (Recall that Valentine's day is coming up!)
• Google Squared: A different way to search.  Try this search of "Fortune 100 Companies".
• WolframAlpha: They call this a conceptual knowledge engine.
• TrueKnowledge.com: Imagine being able to ask questions such as "Give me a list of all the countries in the Airlines Industry around the world.  This web site isn't great, but it points out possibilities.
• Yahoo Pipes: Interesting interface for building queries.  What if you could query the SEC database like this. Try running some of the sample queries, or be bold and venture to build your own query.
• OpenAnalytics: This is just a sample, but imagine the possibilities of this interface.  Try one of the sample companies.
• 15 Stunning Data Visualizations: This is truly stunning.
• Bubble Charts: Hans Rosling analysis of data using bubble charts.  This is a 16 minute video, well worth watching.
• FDIC Bank Database: This isn't bad, wish the made the data available in XBRL so you could query it from Excel.

One day!

Malcolm Gladwell, author of Blink, helps one understand variability and how grouping things into clusters can be used to better understand which is the best pickle and what the best spaghetti sauce is.  The answer is that there is no best pickle or spaghetti sauce.  But there are best clusters of pickles and spaghetti sauces.  Watch this TED video to understand how to meet peoples needs.

So, I am sure you are asking what the heck this has to do with XBRL.  Profiles of XBRL is basic a clustering of user needs.  

There is no one right approach to using XBRL.  It is also true that each system making use of XBRL and inventing their own specific approach to using the general functionality which XBRL provides.  Each system creating its own architecture is simply too inefficient.

The middle ground is to cluster needs of a system. Is extensibility needed? If yes, then you take the pieces from XBRL which cause the least issues relating to extensibility.  You wrap your set of characteristics into a bundle, that is your profile.  Or, rather than building your own XBRL profile, even better is using someone else's XBRL profile which is known to work well.

Watch the video and you will better understand how clustering user needs into profiles can help you see XBRL just a little differently.