I think it's interesting to compare it to the family trees of Indo-European languages -- it's a very different shape, natch. There's lots more cross-fertilization in programming than in natural language. Of course, one has to wonder how much of the cross-fertilization in [human] language change goes ignored in plots like these approximations of language histories.
Another nice feature of the O'Reilly poster is that it shows the duration of various [programming] languages -- most of the old ones (Forth, anybody?) have petered out, but O'Reilly shows the sheer persistence (for example) of C. Some family trees of programming languages don't do this nearly as well, though to be fair that's not necessarily what they set out to do.