You can certainly have a decent idea of what information is there by just looking at the bytes and reverse engineer the protocol. This is how unix pipelines are composed as a matter of fact.

Now, of course it would it be better if each program formally described the schema of its output and the system provided tooling to act on this information, but being able to string together uncooperative programs via their common denominator (i.e stream of bytes), is powerful and the unix way.