How to make great software

It's just a spreadsheet, but it's been programmed to answer, with startling accuracy, questions like, "Is Iran going to build a bomb?" and "Which financial firms will commit fraud?" It's a fascinating piece of software, but this quote underlined a critical point:

I think the model is, I’m sure, brilliant. But lots of other people are good at math. His gift is in interviewing. I’ve said that flat out to him, and he’s said, ‘Well, anyone can do interviews.’ But they can’t.”

The software's creator performs the interviews to generate the inputs for each prediction.

As developers, we often get caught thinking it's the technology, it's the algorithm, it's the technical expertise that makes or breaks our software. Sure, that's important. If that stuff didn't work, it'd be useless.

But that's not what makes it great. It's talking and listening to the users and finding out what they really want. It's understanding the domain space so well that you pull in the important variables.

Anybody can do that stuff. But, as is constantly proven, they can't.