Similar to what others have said, I don't think we can have much depth to a discussion of your number one; "The best possible course that could be built given the constraints of the property, the purpose of the course, and the client's wishes." The only hope we have of making any comments along those lines is if we have the principals of the project, architect and developer, in the discussion to give us clues as to what the goals of the design tried to achieve in terms of what market the project was aimed towards, what style and strategy was attempted to be presented, what land specific issues were present, and what the budget was. One reason I thought all the many discussions and comments were good regarding Dismal River is that we had the architect and principal developer of the Doak course in the discussion and being quite candid about these issues that most would not otherwise understand just as a casual golfer-user-observer-rater. Combine that with the very superficial understanding the hobbiest has of the actual usage and capabilities of the equipment and supplies used to construct the course on a given type of land, and we are mostly in the dark and shooting from the hip on these matters.
Yes, as hobbiests, it is all about comparison to other courses we have seen. It is helpful if the hobbiest has done as Tim Weiman describes (and many of us fanatics have done) and go out and follow the construction process and engage those working in the ground with questions and observe the progression of how a project comes together. Sometimes, the gist of our golf archtecture discussion threads gets down to many just declaring all the golf courses on their belt notching list that they have played, and not much about the actual merits of the design-construction. But, as hobbiests, we have to rely on other courses we have experienced to give any context at all to the validity of our discussions.
On the third aspect, of course it generally comes down to personal taste and standards. Some aspects of evaluation are somewhat consensus ideals (see other thread currently being discussed). It seems if you take a focused group like GCA.com and spread out discussions over a long period of time, a few 'ideals' start to form up and build a consensus. But, not universal ideals. Just, consensus if you think somewhere north of 50% of the group feel a certain way about design characteristics and merits of a particular course presentation.