Comment by lusr | Hacker News Reader

Comment by lusr | Hacker News Reader

Parent Full thread

Yes, I believe that images (encodings of electromagnetic radiation of particular wavelengths) and audio (encodings of mechanical waves of pressure) are universal concepts aliens are also likely to be familiar with, whereas understanding the information encoded within the transmission requires interpretation within a context we are highly unlikely to have here on Earth.

This is just my intuition and speculation but I would guess somebody could formalise some support in terms of Kolmogorov complexity (http://en.wikipedia.org/wiki/Kolmogorov_complexity): there are only so many ways to represent an image or audio, and over time an intelligent race will discover the least complex method and use it. Therefore, alien races of equivalent levels of technical advancement in a given field should be able to decode raw (uncompressed and unencrypted) transmissions of common types of information (images, audio).

Any additional complexity on top of this method will probably be quite structured, e.g. with television signals we have portions of the signal that are blank because during that time the CRT of old televisions would be retracing and not displaying data. (Actually, I believe these "dead" portions of the video signal are sometimes used to send additional structured data, e.g. Teletext, which would make the structure of the signal even more apparent.)

I suspect it'll be much easier to look at alien video transmissions and figure out what they're about than trying to listen do the same with audio transmissions because video would provide objective context (assuming it's not just a video of talking alien heads, or whatever the equivalent of a "head" for them would be).

With respect to encrypted transmissions, if the entire protocol is encrypted from start to finish, and you're using the right cipher* then no, you can't distinguish encrypted data from random noise. This is the principle behind TrueCrypt's hidden volumes (http://www.truecrypt.org/docs/?s=hidden-volume).

* What's the right cipher? So far no nobody has found a distinguisher for AES, and that's in common use, so that's probably good enough to make an alien race blind to us; i.e. we could be picking up signals from a race as sophisticated as c. 1998 humans (http://en.wikipedia.org/wiki/Advanced_Encryption_Standard) without realising it.

Note that encrypted traffic, indistinguishable from noise, is almost always then encoded for transmission in a format that contains redundancy (frequently, there are more than one such layer).

For example if you electrically sniff an ethernet cable over which only SSL traffic using AES is being sent, the first thing you will notice is a concentration of energy near 200Mhz. Then you may see bursts of a Manchester-encoded signal; further study reveals the regular structure of Ethernet headers and trailers, along with IP, TCP and SSL headers. Only within that will you hit a brick wall at the AES data itself.

Even if the payload is encrypted it should be possible in most cases to detect a modulated transmission.

That's a good point. In effect, if we look at this from the perspective of the OSI model, you're saying that whereas the application layer may contain encrypted data the lower layers may contain structured data. So the logical question that follows: is there any reason to believe aliens might use encryption all the way down to the data link layer? Maybe our neighbours are all at war and need to disguise and protect signals from each other?

Have you read ithkuil's comment? http://news.ycombinator.com/item?id=4188160

Yes, but spread spectrum modulations are very much a minority among all our transmissions, so I see no reason to expect any different from other civilizations.

Today yes. But imagine that our technology gets refined in the following 10k years, even assuming no more technological breakthroughs, but only incremental improvements.

Even today we use some technology which is not really necessary to solve a given problem, just because it uses off the shelf components which are cheaper even if more complicated (e.g. a embedded cpu instead of dedicated and simpler electronics)

> then no, you can't distinguish encrypted data from random noise.

But, would there not be a noticeable difference between random noise in space, and that same space PLUS someone broadcasting random noise? (And would it depend on whether or not they cared about being heard, even if they still wished their transmission to be encrypted).

It just seems to me that if there is a particular point that is continuously broadcasting randomness, would it not be noticeable? Or is what you are saying that that is essentially what stars are doing is broadcasting randomness, across a broad spectrum, and we can't distinguish between a star and an intelligent race....they are literally indistinguishable.

Yes, as I understand it (I'm no expert though), we can't distinguish between the background noise of the universe and an encrypted alien signal. I believe the main problem has to do with all the various possible sources of background noise constantly moving and interfering with each other through the galaxy.

That being said, from a physics perspective perhaps there's some way to say "these bits of random noise are unusually powerful, and based on our multiple noise detectors over a large area we've managed to estimate the path these signals travelled and there's only a planet on the other end with nothing else appearing in in the way (over the course of time it took for the signal to arrive) to explain them as mere background noise."

Reply on news.ycombinator.com