The commentaries posted to NB and your presentations should include subjective (e.g., evaluative) reactions. The syllabus lists these types of contributions: (a) compare the work to related work; (b) hypothesize about ways in which the work could have been improved; (c) think about ways to expand on the work (conceptually or computationally); (d) critique the work, including its conceptual framework, methodology, and/or results; (e) think about how the work could benefit NLP applications or other NLP tasks.
Though your comments may be subjective, they should not be rants or flames and they should not be insulting or sarcastic. You should use the type of style appropriate for a blind review of a paper.
Throughout the course, please try to balance the polarity of your comments. As Mitzenmacher and Ramsey 2000 discuss, it is always easier to tear down than to build up. What are the good ideas in the paper? (See the "Read Creatively" bullet in Mitznmacher and Ramsey 2000). In other words, please do not make your reactions (in commentaries and presentations) composed entirely of criticisms of the paper. Those are valuable, but it is perhaps more valuable to find something interesting in a paper and think about how to extend it, tie to other research, link it to a larger issue, use it in an application, and so on.
In your subjective reactions (both in commentaries and presentations), go beyond the paper itself in one of the ways listed above. Please be sure that your reactions are sufficiently substantial so we have the basis for fruitful discussions. A fairly boring comment is simply stating that you would like to have seen another analysis or another experiment. Why do you want to see it? What interesting hypothesis do you have r.e. what would be learned from that analysis or experiment?