A place to share ideas, thoughts,
best practices, and questions about KM in
a government environment
Help me design a multi-tenant government data dictionary
October 21, 2010 at 5:27 pm #113262I want to design a data dictionary for government, shared by multiple agencies, that has fields in it that facilitate discovery of sharable data across agencies.Since multiple agencies would share the same database, they would be able to find data integration opportunities where the data is related to similar lines of business or functions.For example, if there was a data table related to parking lot management for state parks, all states using the data dictionary could more easily find sharable apps or data integration opportunities with other agencies using the shared data dictionary. The dictionary could foster collaboration by linking to discussion forums.Can you help me design this multi-tenant government data dictionary?What would the fields look like? How would metadata aspect of it be designed?The federal government has already categorized all its functions with standardized reference IDs:A good candidate for a field in the data dictionary would be the LOBs (lines of business) in the federal Business Reference Model (BRM), Technical Reference Model (TRM) and Service Reference Model (SRM).Example of what the dictionary might look like:DataDictionaryTable• AgencyName (E.g., Dept. of Agriculture)• DataFieldName (E.g., resumeLink)• DataFieldDescription (E.g., hyperlink to employees’ resumes)• DataFormat (E.g., 40 characters, alphanumeric)• SRM (E.g., 75.752.611 Resume Management seehttp://www.whitehouse.gov/sites/default/files/omb/assets/fea_docs/FY10_Ref_Model_Mapping_QuickGuide_Aug_2008_Revised1.pdf)• DataSource (E.g., personnel department)Does this already exist elsewhere? What is the closest thing to this in commercial data dictionaries/metadictionaries or in enterprise architecture discipline/data architecture?Any ideas or suggestions are appreciated.Thanks,Alex GlarosCenter for Government Interoperability
October 21, 2010 at 6:32 pm #113270
Your request sounds like it’s a very large endeavor! A couple of things I would clarify:
1. Data Dictionaries at the Enterprise level typically fall within the data domain especially in frameworks such as TOGAF. http://www.opengroup.org/architecture/ There is more help there.
2. Have you spoken with Naomi Caietti from the State of California? She is on this website listed under the Enterprise Architecture group. She may have some good insight into this.
As for building an Enterprise data dictionary/database, I’ve attached a document that speaks to Architecture principles for data domains from the point of view where I work. This was developed by our Enterprise Architect last year. It can be used for an agency or enterprise and is very generic.
One final point: This is a very large effort that requires a lot of cooperation, communication, and collaboration. I’ve worked on an Agency database that took years to develop and finally implement because of politics, changing requirements, and just plain head-butting.
Feel free to respond if this helpful.
October 21, 2010 at 7:11 pm #113268
The architecture principles doc has some good data governance ideas.
If you know of anyone that can summarize the data architecture components of TOGAF as they relate to my project, please let me know.
The data dictionary would be for all federal/state/local agencies. I’m just creating a prototype and will contact Naomi when there is a working example to demo.
October 22, 2010 at 2:15 pm #113266
Is there any reason that you did not cite the DRM part of the FEA? The model suggested there has a bit more semantics than a traditional data dictionary (but probably not enough to ensure easy interoperability.)
Taxonomies, thesauri, classification systems, synonym rings are all examples of controlled vocabularies.
Without things like common taxonomies people’s terms and data definitions will be quite diverse. So one needs to describe the basic steps in controlling vocabulary and how these steps map to the terminology.
There are various work that provides some basic guidelines and recommendations for creating controlled vocabularies and leveraging them effectively,but looking at the DRM might be a good start.
October 22, 2010 at 6:36 pm #113264
Thanks for bringing up the DRM; I should have mentioned that the data dictionary app is intended to implement DRM concepts.
The DRM intentionally does not delve into specific models or artifacts that the architects will create. I’m creating a prototype app to see if the data dictionary concept works. The least that it can do is offer a free or almost free data dictionary for small government organizations that can’t afford one.
You are correct that some sort of semantic and taxonomy processing features must be added to the data dictionary. I am also planning a data harmonization component.
If you have any ideas regarding collaborative data standards, semantics, etc., please forward them to me.
You must be logged in to reply to this topic.