Computation speed is only one of many criteria. It can be important. We have R code that takes hours to run, and for another paper it takes 5 minutes for the R code every time we change something. Speed is certainly important but not the only criteria.
I agree, it would certainly be useful to compare the most useful libraries, for economists dynlm is fantastic in R. But we only have so many words in blog piece like this, and we didn’t have space
I stand by the historical baggage. If you compare R (and the others) to Julia, there are so many bizarre and inconsistent language features they carry with them. Julia doesn’t have a lot, it might have made bad design decisions as Yuri argues, but it’s not historical baggage.