Hyperslab: reading parts of a dataset

[carstenbauer]

julia> using JLD

julia> save("test.jld", "x", rand(ComplexF64, 4))

julia> jldopen("test.jld") do f
           try
               println("reading everything")
               read(f["x"])
               println("worked")
           catch
               println("failed")
           end

           try
               println("reading parts")
               read(f["x"], 4)
               println("worked")
           catch
               println("failed")
           end

           try
               println("getindex parts")
               f["x"][4]
               println("worked")
           catch
               println("failed")
           end
       end
reading everything
worked
reading parts
failed
getindex parts
worked

Would it make sense to also define read(dataset, indices)?