Grutter v. Bollinger (Michigan Law School):

So far I've only read the law school case that O'Conner wrote. It seems largely an affirmation of Bakke.
The Justice really sticks to the same nutshell that we get from that case- no quotas, but a goal of diversity is OK. Indeed, the first legal analysis segment of the opinion is an extended brief of Powell's opinion in Bakke. To wit, O'Conner uphold's Powell's notion that diversity can serve as a compelling government objective, in no small part because of the longstanding notion of academic freedom--courts ought to give deference to a school's decisions of what makes for better education.
That said, the equal protection argument runs through the the two prongs -- 1) compelling interest that must be 2) narrowly tailored to that interest -- very tightly holding the hands of the Bakke decision.
"The freedom of a university to make its own judgements as to education includes the selection of its student body." says O'Conner, quoting Bakke at 312. That freedom being available, O'Conner is able to list off all the benefits that Michigan and the famous Amici Curiae listed; my law school mentioned the bringing up of the future leader class, several CEOs wrote about the benefits of diversity; as well as the most famous friend of the court brief from the military officers.
Prong two asks whether the school has sufficiently fine-tuned the policy. The first requirement here is that the program not be a quota system. A quota is a "program in which a certain fixed number or proportion of opportunities are 'reserved exclusively for certain minority groups.'"
O'Conner then discusses the same Harvard goals-oriented program that Powell used in Bakke as an example of a permissible policy that is not a quota or set-aside. While such a program (in Michigan, a goal for a 'critical mass') does consider race- that consideration is a flexible one...indeed, that flexibility is one of the crucial elements that prevents the program from being deemed a quota.
Another major factor for the court is that the applicant is evaluated as an individual rather than as a member of a particular group. The later classification lends itself to acceptance based on a "single 'soft' variable."
Quoting Bakke, O'Conner notes that, "like the Harvard plan, the Law School's admissions policy is flexible enough to consider all pertinent elements of diversity in light of the particular qualifications of each applicant, and to place them on the same footing for consideration, although not necessarily according to the same weight." Grutter, quoting Bakke at 317. The point that O'Conner wants to drive home is that the Law School notes the race of the applicant, and does so with the goal in mind of a critical mass, but the racial status of an applicant is simply a factor counting to the individual nature of that applicant.
Finally, narrowly tailoring does not require that the government use every alternate facially neutral avenue available. Rather, there must be "serious, good faith consideration of workable race-neutral alternatives that will acheive the diversity the university seeks." O'Conner lists some proposed alternatives (a lottery and lower admission standards) and rejects them--the school would have to sacrifice its "educational mission" of being a highly selective institution.
Of note, the court rejects the plan that the White house offered--to admit a certain percent of students from every high school in the state. Such a plan "may preclude the university from conducting the individualized assessments necessary to assemble a student body that is not just racially diverse, but diverse along all the qualities valued by the university." It seems to me that O'Conner sees in a percentage plan just the kind of group-admissions rather than individual-admissions that she finds bothersome in quotas. I had not thought of that argument (I had only dismissed the percentage plands as being overly-reliant on segregated schools).

Finally, comes the suggestion of a sunset on the program upheld; what will certainly become a famous bit of this opinion-- the 25 year provision. O'Conner projects that in 25 years this type of policy will no longer be constitutional permitted because it will not be sufficiantly needed in order to acheive the goal of diversity. As of now, I assume this is part of the second prong analysis- that the program is norrowly tailored. Justice Thomas attacks the argument- saying that what is constitutional in 25 years is constitutional now. But he seems to miss the logical boat here. O'Conner makes no suggestion that the Constitution will change. Rather, she guesses that in 25 years, a conscious look at a person's race on an application won't be needed in order to fill the diversity goal. It is a sociological prediction about the context of this policy in the future, in that such context is a key factor as to the reasonableness and the need for such a program.

More later.