The r.http command was controversial, even internally. I had to call in a lot of favors to convince people to get it in, but here is why I was convinced it's a good idea:

  - It fits! It just seems to work magically well with the rest of
    ReQL on so many levels! JSON fits, streams fit, lazy evaluation
    fits, even batch prefetching fits! Everything works so
    wonderfully and provides such a great interactive experience,
    it almost would be silly not to add!
  - It makes a use case that's really important to me
    personally (adhoc analytics) an order of magnitude easier. Over
    the past few months I ran into a few other people that also do
    a lot of interactive adhoc data analysis with Rethink, which
    redoubled my resolve to add r.http.
  - It makes example datasets *much* more elegant. Reading
    RethinkDB docs? Just call this command to get the dataset in so
    you can play around with a command. (We haven't updated the
    examples yet, but we will). That makes the first experience
    with the product better, and makes the learning experience much
    more pleasant.

So I hear what you're saying, but I respectfully disagree. If you think of it as a signal, think of it as a signal of deep passion for the product and our users! Ornamentation for the sake of ornamentation isn't what we do.

(On an unrelated note, I really love your use of 'frippery' -- I'm a bit of a word nerd, and this is a really neat word!)