Reverse Engineering: Black Boxes 101

All scientific research taken as a whole is an attempt to reverse engineer the universe.

Reverse engineering, as you might suspect, is the reverse of engineering.

Engineering is the art and science of creating something which performs a function: you need to hold up the weight of a bridge so you engineer a column with the right width and depth to hold the weight you need at the proper height.

Reverse engineering is the art and science of determining the structure which performs a function: you see the the bridge, know that something needs to hold it up, then look below and lo and behold there's a column. ( If you were monkey-wrenching, you'd then proceed to blow it up and see if the bridge comes down, but that's another post. )


There are two flavors of reverse engineering -- 'open box' and 'black box'.

Open box reverse engineering is where you can see directly see the structure, either by looking directly at it or by having some other kind of direct perception -- a microscope, radio telescope, fMRI -- you get the picture.

In this kind of reverse engineering, you can just observe whatever it is in action and if necessary do some simple tinkering directly with it's structure to figure out how it works.

Black box reverse engineering refers to working with structure you can't see and have to work with it indirectly. (And this, my friends, is where it starts to get really fun. )

The academic black box is just a black box with any number of inputs and any number of outputs, usually pictured like this:

Inputs ---> [Black Box] ---> Outputs

In this post, I'm going to cover some well known black boxes which have only outputs. I'll cover inputs in a later post, because my main objective here is to give a clear idea about what a black box exactly is.

Astrophysicists mainly work with black boxes simply for the fact that most of what they study are at least several light years away and so there's no way to get to what they're studying and even if they did, it's impossible to get a look inside something like a star or a black hole.

Instead, they study the outputs of their black boxes, outputs like the qualities of light coming from a white dwarf and the gravitational pull of a black hole.

( Actually, in the second case, it's a long chain of outputs. Since they can't measure the gravitational pull of a black hole they measure the mass and speed of the stars orbiting it. Since they can't directly do that, they measure the qualities of the light coming from those stars. The chain looks like this :

Light from the stars indicate
-- their mass and speed of orbit which indicate
---- the mass of the black hole [box] which indicates
------ something about the structure of the black hole itself )

In other cases, like the study of geologists, the black box doesn't exist as one whole thing in one place but as a series of processes over time, but luckily the outputs (rocks) are easy to study. For example, a geologist can look at a rock and tell you it was initially formed in a volcano, then moved by glaciers where it wound up under the sea for a few thousand years, finally to find itself in a desert, where you found it and brought it to him. ( If there's anything that geologists like to see it's not a new kind of rock, but an indicator of a process they haven't seem before. )

My favorite black box is the human mind. Everything someone says or does is an output, and as Banachek pointedly pointed out: "you can't not communicate". This leads into some seriously fun 'lesser magic' involving outputs few people pay attention to ( not body language, but that's fun too. ), but I'll have to leave that for another post.