So, you have to be able to surface constraints and show them to people. Now, the report shows the data well, but it typically does not show the constraints that we’re applying to retrieve that data. This often leads in the BI world to problem solving for us as the BI professionals supporting others. These aren’t the constraints you’re looking for. You have the report using these constraints, but you left out some data. That’s why your numbers don’t match.
Well, you input those things, and it immediately pops back out and changes the set of information. That has to happen very quickly, which is hard if you have to go through this design/execute/view model. So, we have these things where we can’t really reduce the analysis challenges using a hierarchical, master detailed, paper based model of looking at things. You need a true interactive model.
Now, all data mining technologies tried to address some of this and were partly successful except that those technologies still had to repeat data, and impose constraints, and they still use a user interface very tightly related to a paper based model. And so, when we look at that, you see we’re still in a publishing world. And the problem is that you’re in that publishing world where you have to build the cubes with the users. And that’s a very bad model because over time things change, and you have to constantly adapt those cubes, and the users can’t do it for themselves.