Personally I think the hardest bits of Comp. Sci. are actually formal-philosophical. A mixture of mathematical formalisation, with philosophical analysis, and theorems that have deeply philosophical implications.
And most of the science bits, in the sense of empirical science, are really about using computers as a modelling tool for x (e.g., society), which isn't comp. sci. any more than astronomy is telescope science. The other science bits around how people use computers, including empirical science around the usability of languages, is partly psychology, and in some cases wholly.
It's a mish mash.