Agreed. IMHO, this is a *major* problem today. Ease of use is being equated to better or more powerful when really it means less and less differentiation between competing products mostly using the same technologies. (I-deas and CATIA being a notable exception here.) I guess if we could say we have solved the core MCAD problems in the same way we have word processing problems, for example, this might be a good thing. I am just not sure we have traveled far enough down that road yet.
It is interesting to note the direction Alias took with their products. They adopted the good parts of the win32 environment while continuing to retain the elements of their GUI that were well thought out and powerful. Alias remains difficult to learn, yet is very easy to use once the learning has happened.
Alas, it is also more difficult to articulate a unique value proposition. These two things combine to reduce much of the discussion to price. While that is a good thing for us buying software, in the end we all lose out on potential niche technologies and options that may hold the potential for competitive advantage over our peers.
Take a look at the I-deas analysis environment sometime. It is different in the way Alias is described above. Has the same advantages as well.
Putting a nice interface on a complex task does not assure the user of anything other than a nice interface really. Complex stuff is still complex.
Your point on the solver interface is an interesting one though...