The Google Books settlement: five brief observations

Some quick responses to the proposed settlement, explained here (PDF):

1. It has nothing to do with periodicals (2).  Since searches of in-copyright periodicals are arguably GB's weakest point now, this is frustrating.  At present, such searches not only produce the infamous snippets-without-relevant information, but also frequently generate results without full bibliographic data (thereby making it  impossible to do an ILL).  
2. For individual users, purchase prices per in-copyright, OOP book would top out at $29.99 (6).  As an academic, that seems reasonable to me, given what OOP scholarly works can run.  Others may disagree, however. 
3. Snippet view for in-copyright, IP books will vanish (5).  What Google gives, Google can take away?  Presumably, anyone who wants to actually see their search results will run over to Amazon, where "search inside this book" already generates full-text results that don't appear on Google.
4. Annotation function (7).  This is nice, but as others will surely point out, there are non-Google related programs out there that will perform similar functions.  
5. Free public access, sort of (8).  If I'm understanding the description of library PAS correctly, students and faculty would actually have to be in the library (or at an associated terminal); it would be impossible to access the system from, say, a smart classroom (unless it was the university's designated terminal, I suppose), a dorm room, or an office.  Moreover, the agreement strictly limits the # of terminals that could host the PAS.  That's not very convenient, or, in some cases, even very useful.  Of course, we could always work around such issues by, yes, paying up.