K:
Yes BPMN diagrams enable capture of INPUTS and OUTPUTS. Unfortunately, they put INPUTS and OUTPUTS and flow of control and sequencing and the kitchen sink all on the same diagram. The result, especially for larger scale efforts, is diagrams that soon get so complex that the analyst can not tell where his/her mistakes are. I have seen this happen in the real world more than once.
The creators of data flow diiagrams where well aware of this problem. That is a primary reason why on dfds the inputs and outputs (typically data flows, but can be flow of material or tooling, etc) are seperate from flow of control, etc: the inputs and outputs are on the high- level diagrams and flow of control, etc. are on the lowest level diagrams. Principle: If the BA wants to successfully diagram anything, he/she must not try to diagram everything - at least not initially. The need is, to the degree possible, start out by focusing just on the higher level stuff and to postpone the detail until the appropriate time.
Another reason why the flow of data is on seperate high- level diagrams with dfds is that systems, especially at higher levels of abstraction, are, as Yourdon (DeMarco?) used to say, very asychcronous. This means that many tasks can be happening at the same time or happening in any order - prohibiting the analyst from documenting them via sequencing/flow of control, exept at the low level. The creators of the BABOK 2.0 realized such. That is why the second paragraph of Section 1.4 in the BABOK contains the sentence "Tasks maybe performed in any order as long as the sequence inputs are avialable." They were wise enough to realize the asychronous nature of systems (in this case the system being the set of interrelated tasks that a BA performs) - and to use data-flow-diagram-like input/process/output diagrams.
FYI: While the BABOK 2.0 employs input/process/output diagrams, it does not utilize integrated input/process/output diagrams (i.e., data flow diagrams). This puts the task of integrating the diagrams within the BABOK 2.0 onto the reader - making the document very difficult from which to get a systematic understanding of its content.
Tony