David Maddox
Sound Science LLC
I participate in lots of discussions about data, and information, and how to design databases. Sometimes a database is just a repository of data – a place to store it for some future use. More often, though, a database is meant to be an active information delivery system. Let’s call such a thing an IDS.
And what’s the most important function of an IDS? It is the delivery of (1) information that users want and (2) in a form that users can use for their purposes. Maybe I should say that this really is the ONLY useful function of an IDS.
So I am troubled when I see databases developed “from the bottom up”, because what this really means is they are developed more with the data in mind than the user. What does the user want to know from the data? In what form do they want to know it? These are the critical first questions in any database development project. They must be asked before any other work is done. Develop from the top down.
Database developers often speak about “Use Cases”. Use Cases are depictions of how database tools will look and function. What are talked about much less often are “Scenario Cases”. Scenario cases don’t describe the how of the data searches. They describe the why. Good Scenario Cases are created from discussions with potential users – focus groups or interviews – that discover what they want to know, what kinds of searches they want to make (search terms, key words, etc), and how they want to see the information summarized and displayed.
Without investing in early focus groups and the creation of Scenario Cases, lots of cool programming based on Use Cases will be wasted and misguided. And much useful information will lie fallow and unused, buried within a database that could have been more useful. That is, it adds up to less that it should, and doesn’t honor all the work and money that was originally invested to amass the data. 1+1+1=2.
