I went to undergrad at Swarthmore College, a small liberal arts school outside of Philadelphia. Swarthmore requires all students to take writing courses in a variety of disciplines. I wrote about contemporary Irish poetry, educational philosophy and practice, and central African history. I also wrote thousands of lines of code and many proofs - writing courses are based on the writing of the discipline.
When I got to grad school I was talking to another new student and discovered that this was not universal. He had taken about 40 computer science classes (some were 1-credit classes) to my 10. He asked me what I had been doing taking anything else - why bother learning about Plato if you know you want to go to computer science grad school? Besides my personal interest, I responded that it helped me learn to write. He shrugged this off saying something about how he would pick that up as he wrote papers.
At the time, my belief in the importance of learning to write - purposefully learning, not just blundering along in hopes of picking it up - was instinctual and ingrained. Now, after reading many poorly written papers and discovering in my own writing how hard it can be to present technical details that you have been living for months in a way that others can understand without having to live the process, my belief is founded in experience. I have read papers where even titles and opening sentences made no sense. The struggle against deadlines does not excuse this, rather it makes my case that we should learn to write before the fact. There's plenty more to learn in grad school.