The risk is to bite off more than we can chew, to establish costly and time consuming evaluations that interfere with instuctional time and actually detract from student achievement, rather than enhance it.

I think that the recommendations are well-crafted in that they are relevant and provide flexibility.  A teacher evaluation system built on these recommendation would allow for inputs in most of the domains that matter, and would provide options regarding what data would be collected.

One risk that I see is related to measuring growth and value-added by a teacher. I think there needs to be a longer view of the student's performance.  If data is available to track student progress over multiple years, information about a student's learning trajectory (across teachers/years) should be taken into account -- in addition to one year gains -- so that even if a student doesn't grow by a full grade level in one school year, a teacher who produced a similar rate of gain or a higher rate of gain over previous years should not be considered ineffective.  I think it is crucial to acknowledge, and not forget, that some significant portion of student achievement is due to student-related factors like intelligence, prior knowledge/learning opportunity, and work ethic/motivation, which teachers have little or no control over.

Another risk -- that applies to all efforts of this sort -- is that ideologies and political influences could lead to a whittling down of the options for fulfilling the measurement requirements in any given domain.  This could lead to factions where one side denounces all standardized test results and another devalues all teacher-derived inputs, although recommentations allow for both to have an input.

The recommendation that the process remain "living" helps to address the risk that early choices will become crystalized by tradition in ways that prevent inclusion of other measurement options.  I think this is especially important in the highly charged and politicized decision-making environment in education, but there is a risk that this could lead to pendulum swings that make evaluation results hard to compare over time.

Overall, though, I think the recommendations begin at a great place that allows specific implementations to be locally responsive, but with elements that make them comparable.  There is also a lot of room for stakeholders at the local level to craft something specific that operates in the spirit of the recommendations.