There are dozens of dictionaries, and while they're relatively small, performing sequential network i/o for all of them isn't particularly cheap.
In parallel, they getting all of them should cost <8s.
FAQ | Showcase | Event descriptions | Configuration information | Known Issues | Possible features | Deprecations | Release notes | Helpful scripts