There is an interesting post at "Leiter Reports" on book reviews. As I've noted before I read a lot of book reviews, and a great many book reviews end up being rather worthless precisely because the author can't say anything dispassionate about the matter. That X does not like Y's book is largely useless information except insofar as it might be taken as evidence of bias, or else insofar as it indicates a recommendation against reading it (and it doesn't take much to do the latter). What a book review really needs to convey is (1) what the book is about, in such a way that I know if I have to read it if I'm researching this or that subject; and (2) what I have to keep in mind or know if I do read the book, either so that I can get the most out of reading it, or so that I can make sure that reading it is not a complete waste of time. Heated rhetoric and vague preliminary comments about incompetence, uninformativeness, amateurishness, etc., just uselessly take up space and tell us nothing; just say you can't recommend that it be read and that it has many problems (which are, among others, such, such, and such) instead of filling the page with negative fluff. This is true on the puff-piece side as well; it does the reader no good to know that X likes Y's book, except where it shows bias or signals whether a person would recommend it (and it takes very little to do the latter).
On tone, I think disputes about tone very rarely, if ever, "mask" disputes about substance, as Leiter suggests (and it's certainly not true "of course"); they are disputes about what Hume calls "the lesser morality," or etiquette, and while secondary to other issues, these issues are not trivial distractions. (They can be abused, of course; in this they are no different than 'the greater morality'.) There are issues of professional standards, reasonable expectations of respect, and the like. The problem with saying that someone's claim is "preposterous in the extreme and easily refuted" is that it is not any more substantial than calling your colleague ignorant and stupid. Should we, as a serious profession, really tolerate things like this? There is generally nothing honest about it, either; it's often a blatant substitution of rhetoric for argument, labels for reasons, and where it is not, it is often otiose.