 |
|
|
|
|
|
|
|
||||||
 |
 |
 |
 |
|
||
 |
 |
 |
|
|||
 |
|
|||||
i have been watching and listening to the blame game (also known as the 9/11 commission hearings) over the past few days. again and again the finger seems to be pointing at the dysfunctional relationship between the cia, fbi, presidential advisors, and the president. many are complaining that there is little communication between agencies and that the agencies are not sharing information.
people are missing the boat on this. lack of communication and the sharing of information are not the problem. those issues are symptoms of an unusable system. the real problem is information architecture, taxonomy, and metadata.
i assume there is a grand data base of intelligence some where. i am certain the information architecture sucks, rebuild it. next, multiple taxonomies are essential to supports multiple users with different views of the data contained. finally the most important piece is metadata. good metadata can allow the users to transend the rigid architecture and taxonomies (and also hide any flaws in the ia or taxonomy). good metadata will allow the user to realize information relationships that have gone un-noticed.
if those elements were in place, the relationship between the cia and fbi becomes less important, they can continue to ignore each other but have access to each other's intelligence in a usable way. sharing becomes a non issue, as long as data is put into the system, it is available via many different paths to many different users.
as an example, in a company like nike, many different people need the exact same information to do their job well. the marketing team, the advertising team, the board of directors, and the ceo all need access to sales reports, but may need the reports for different reasons and/or presented with a specific spin to be useful. each group will also have it's own lexicon, assumptions, and understanding of sales reports. it is essential they are provided with the data requested regardless of how they ask for the reports.
the users of the system just need the information, they shouldn't have to know how to ask for it and be denied because they did not ask the system for it properly.
over at airbag, greg has a great post with similar thinking but focused on the visual design of pdb documents. go check out his pdb redesign. it is an amazing improvement (done with no user value investigation, mind you!). in his comments, greg ponders the possibility of a grass roots, open source, movement to get the government's shit together. interesting.
all that being said, i could be wrong. my expertise clearly lies in poo humor.
Posted by griff at April 14, 2004 01:09 AM