[btrettel]

I don't think academics are incentived to publish truth in general. In my field (fluid dynamics), researchers seem to focus on producing as many papers as possible that sound good but may have problems beneath the surface. This may be different in math, because the standards are much higher.

The distinction between whether science or math creates or discovers truths is philosophical and not something I'm that interested in.

My goal is to publish papers as likely to be true or useful as possible. So thinking about this in terms of "product development" is really just about increasing the chance of discovering the truth. (Or developing a useful model, which may not be strictly true but is a good approximation of the truth. As an engineer, what I do is more likely to be in the latter category.)