I could write a lot on the topic (and hopefully will some day), and have written a bit already (in particular my essay "Death of a Discipline" in the spring 2014 issue of
differences, there regarding the Digital Humanities) but basically it's contentious because it suggests that there is some set of practices/systems called "languages" that includes "natural" languages, programming languages, and/or formal languages, and that this category called "language" is one about which generalizations can and should be made. Outside of some very abstract ones, however, that just isn't true, and those generalizations that are true tend to be characteristics of formal or programming languages that were designed to be like natural languages, and appear a fortiori in (some, or possibly all) natural languages.
It is culturally contentious because it leads to the widespread and mistaken view that "programming languages" and "natural languages" are the same kinds of thing, which I believe is demonstrably and utterly wrong. One piece of social evidence for its being incorrect (which I hope any linguists on the list will feel free to weigh in on, but as I dabble in linguistics a bit it is a generalization I am fairly certain about) is that we have a field that studies what it generally calls "language," namely linguistics, and that field requires its scholars to become experts in a language, and that is always a human language. So specialists who use the naked term "language"
always mean what we mean by "natural language." So far as I know, nobody has yet gotten a PhD in Linguistics where their specialty was Ruby on Rails, and I hope it remains that way and believe it will, though at this point one never knows. In part this is because linguists have some nice criteria for what qualifies as a "language," and programming & formal languages lack many of the most important characteristics. That's why in scholarly lists of languages, only "natural" languages appear (one of the most widely-cited and authoritative of these lists is the
Ethnologue:
http://www.ethnologue.com/browse).
Yet we see everywhere a great deal of evidence that many, particularly those strongly identified with computers, believe in the existence of this super-category, often suggesting, for example, that students should learn programming languages instead of or as a reasonable substitute for learning natural languages. They are just different sorts of things entirely. The argument whether students should learn to program is a separate argument from whether they should learn foreign languages (at least in so far as any such pedagogical questions can be separated). And note that there is a political-cultural dimension here as well: it is always (at least in my experience) computer advocates wanting to claim that programming languages "are" languages, and never the other way around (ie, not linguists wanting to claim that programming languages are the same kind of thing they study: which is not to say they don't have linguistic features, as of course they are designed to do, but that "hey, those programming languages are languages too" is a claim that only emerges in one direction).
as long as I'm writing, I greatly appreciate Pierre's citation of the recent D. Nofre, M. Priestley, & G. Alberts paper. I had not run across that yet and, on quick glance, it looks like both an excellent analysis and to be an invaluable source of information, and adds a great deal of context to what I mentioned metonymically as the work of Grace Hopper and "others."