Which is a fair point - I grapple with how much of what I do is science, how much is engineering, and how much is neither. But, the problem is, that's what the word "computer science" means now. We could try to retroactively redefine a word to fit what we feel is a more natural division, but how often is that successful?

In what I do (mostly business cruds) it is mostly an art. The skill is not to get it done, but to organise the code effectively so that it will be easy to debug and adapt and reuse etc. It really doesn't feel like a science and more like being an editor of an anthology.

I consider that software engineering. "Doing" computer science means you're doing research. To keep abusing the physics analogy, your work (software engineering) is to mine (computer science) as mechanical engineers are to physicists.

Interesting. I talked about this with a friend, who did her BS in Math and Physics and now is a PhD in engineering. What we finally agreed was that physics is just so vast, that the `practical' aspects have spawned engineering branches. i.e mechanics == mechanical engineering, electromagnetism == electrical engineering, atomic physics == nuclear engineering and so on. Maybe we should do the same with Computer Science