My experience does not agree with what you wrote in your last paragraph. Writing code as a computer scientist is equivalent to doing lab work as a biologist or to building instruments as an experimental physicist. You can't do research in CS if you don't get your hands dirty, unless your research in theoretical computer science. In exchange for this, many doors open to you after you finish your PhD that would otherwise be closed (unless you did your research in theoretical computer science). And among the best and the brightest that I met in industry, many are in fact PhDs.