I agree that writing skills are important, but I also think that strong mathematicians are skilled writers as well. One general drive in mathematics is to lay out complicated ideas in ways that make the resulting theorems and questions seem completely natural and even trivial. It's no small task, and in my experience it has drastically improved my technical writing skills (where before math I only had historical/literary/debate writing skills).
> One general drive in mathematics is to lay out complicated ideas in ways that make the resulting theorems and questions seem completely natural and even trivial.
.... to other mathematicians. And this is often a problem. I've had to fire mathematicians-come-developers who thought that writing their specs like a maths paper was a viable way of communicating their ideas, because they proved totally unable to communicate their ideas in a way that the rest of the team understood.
Lets be clear: I'm not saying all mathematicians do this - one of the brightest and best developers I've ever worked with was also deeply into maths yet is also great at communicating.
I just don't for a second believe that being a mathematician in any way implies that you'll necessarily be skilled at writing for a non-mathematician audience.
Mathematical writing is a particular genre of writing that doesn't work that well for technical and design documentation. This is not to say that these things can't contain mathematics, but you can't follow the usual template of:
* Abstract & Introduction
* Lemmas, Theorems, and Corollaries, each with their own proofs
* Some examples sprinkled throughout
* Maybe a concluding section with further questions or areas of investigation
