I've found is that there's a fundamental mismatch between the people who specify what they want in requirements, in terms of software, who can't really articulate it. Who say, "I kind of sort of, I know what I won't want. Just show me something of what the software might do, and then I'll tell you what I know I don't want it to do." Is there a fundamental mismatch in the way that people who are specifying requirements, and the people who are implementing speak?
Mauricio: Well, I think that's the main analysis problem. In Brazil, we have an analogy. I don't know if that will work in the States. But it's something typical from my country, which is the analyst cannot be a waiter. Like, you can't just go to your customer and say, "What do you want?" And the customer says, "Well, I want a steak, and I want some ice cream on it." And then you write it down, and you bring it to the customer. And then he says, "Well, I can't eat this." And you say, "Well, that's just what you asked." That's the waiter analyst. No judgment at all, right? But the analyst cannot be a doctor, either. Like, the client says, "Well, I want a….." "No, no, don't say anything. I know what you need." And the analyst prescribes what the clients will have. That doesn't work, either. The analyst should be like an architect, and that's something that allows for creation, allows for being sensitive to other people's needs, you know. And that's the image you should have. This communication problem should be minimized by an analyst who knows what he's doing. And who is able to hear what the client needs, too.
Carol: Do you think there's differences in… Well, I will finish my question when we get back from break. But I have a couple more questions, just in terms of what Dilbert means to us from an educational point of view, as a potential, and also what ………. And we'll be back.
Welcome back. We're into just about our last segment. And I'm sitting here with Mauricio Aguiar, who works for CAIXA Economica Federal, a leading Brazilian government bank with over 2,000 branches. And he's here for a conference on software measurement, run by the international function point users' group. And I'd like to welcome him back to the show.
Mauricio: Okay, Carol. Here I am.
Carol: And we have time for probably a couple more questions. And then that's pretty much rounding out our entire time. We've kind of talked all over the map, kind of about Dilbert. Dilbert the comic strip that Scott Adams introduced a number of years ago, which has now expanded into books, tapes, and a full-length TV show. And it has made its way down to Brazil and become a dominant force, as much in the IT industry down there as it is up here. And a final couple questions I'd just like to ask you, Mauricio, is your views on how could Dilbert potentially be used as an educational device to kind of change… With your background in neurolinguistic programming, what could Dilbert do to improve the software industry?
Mauricio: I think that the Dilbert strip could be used as an educational device for managers and employees in the IT industry. I don't know if Scott Adams ever thought of that, but the way Dilbert uses language and situations is very, very important for conveying the true meaning of certain interactions between managers and employees. I think managers can get a better