What mainly separates computer science from traditional engineering is the involvement of classical physics.

A more pertinent question would be, what separates art and science? Artistry is in the intuition and science is in the intellect. An intuitive scientist is the groundbreaker. If you seek to be one who brings a new frontier to humanity, do not disregard art nor science, but bring them to bear. The intellect is present to serve as the keyboard for inspiration.

Computer Science has existed in many school's curriculums much longer than Computer Engineering.  This doesn't mean there wasn't any Computer Engineering gettng done.  On the contrary, it was getting done my Computer Science graduates.  Certain instituitions have even attempted to combine the two disciplines and offer CSE degrees.

I'll say that the statement from McConnell "Science programs prepare students to conduct research; engineering programs prepare students to participate in industry.” is simply false, there are ton's of scientists in both research and industry.

I looked around at some schools to see if they have engineering in it.  See below.

Schools:

http://www.cs.ucla.edu/csd/academics/undergrad_program.html

http://www.usc.edu/dept/publications/cat2008/schools/engineering/computer_science/undergraduate.html#bachelor

It looks like UCLA has it with engineering and not.  USC has some courses also.

http://www.eecs.mit.edu/acad.html

MIT is a combination but a very hard site to look around.  Maybe I am not in their league to use the site.

http://bizdiv.fullcoll.edu/cis.html

http://www.csulb.edu/colleges/cba/is/course-descriptions/

Both Fullerton and Long Beach have no involvement with engineering. 

Just a thought, over generalization without careful examination could risk the opportunity to find the true gem. Regardless the scientist or the engineer type, we need to examine the statement in 360 before we allow it to shape our field of view. Sometime in life the field of view reduction is needed to stay focus. And focus is often the key to discovery. hmmm... sounds like a paradox a bit...

As for engineering and science, I think there is very little distinction between the two in the undergrad level. You just have to know all the basic building blocks that you might use in the future. As for graduate level work, it's a lot more theoretical in CS than CE.

In CS, you would need to understand the optimization theory for compiler design that might never get used in the real world. In CE, you would never touch those things. You would probably use some compiler to work on some device protocol leveraging mostly known theories. 

I think in US, often the thinkers are brought down to earth by the real engineering problems. Like how do you calculate the minimum energy orbital path for the lunder lander. It's not as trivial as you might think, but it has been done before by a Caltech mathematician in the Apollo program. That mathematician happens to be my high school teacher. McConnell's book is trying to say you need to be a practitioner to be an American Scholar. I would say you need to be a great thinker, then you can be a great practitioner because you have the pieces to the puzzles. If you don't try to solve the puzzles, too bad. Someone else will come along to solve it.

You can be a great thinker, but if you don't express it, no one would ever know that you have that knowledge. That's why you would need to be a doer. That's why I think McConnell and others got the idea reversed. A great American scholar is a great thinker who is expressive, not a doer who thinks occasionally. At least that's what I have found... A lot of great work are done by mathematicians and physicists who later on do the amazing engineering work/discovery because they have the pieces of the puzzles.

Again, this is just my twp cents. It might not worth much.