An Interview with Brian McConnell04/16/2001
oreilly.com: Why did you write this book?
McConnell: I've been fascinated by telecommunications since I was a child. I am still impressed by the eighth wonder of the world, the public telephone network. Push a button, and you can talk to someone on the other side of the world. Neat stuff, and it gets better every year.
SETI (Search for Extraterrestrial Intelligence) represents the frontier of what is possible in telecommunications. We know the physics of communicating across interstellar distances are no different than the physics of making a cellular phone call. So, we know it is technically possible to communicate with other star systems. The real question is whether there is anybody else to talk to.
SETI is often misrepresented as a narrow discipline, a specialized form of radioastronomy. This is far from true. The search for extraterrestrial life, which is only now gearing up in earnest, touches many different fields of study from biology to linguistics. Unlike many science programs, SETI has captured the public's interest, with over three million people running U.C. Berkeley's SETI@home software on their personal computers.
What's interesting about SETI is that it has the potential to affect our society in profound ways, whether the search detects something or not. If we fail, we'll still advance the state of the art in communication and computer science. SETI@home proved that distributed computing can work on a really large scale. If we succeed, we might be exposed to information from ancient civilizations. The implications of that would be pretty amazing.
We could find something next week, or we could find nothing after centuries of searching. Nobody really knows what, if anything, we will find, but we will learn a lot in the process of looking for other civilizations. Even if we find nothing, we will learn something important, that communicative civilizations are very rare, and we will also develop improved computing and communication technology as a result of the search.
oreilly.com: What makes this book unique?
McConnell: Plenty has been written about SETI, but most of the books and articles in circulation have been written from an astronomer's perspective. My background is in telecommunications and software development, so much of the book is written from that perspective.
oreilly.com: Do you think SETI will succeed?
McConnell: I don't know.
Whether SETI succeeds will ultimately depend on whether intelligence has long-term survival value. I think there is a growing body of evidence that life will develop in most habitable environments, those with active planets or moons with standing water. What this means is that bacterial life is likely to be commonplace on other habitable planets. Animal life evolved from this foundation, and so therefore, complex life will have had an opportunity to evolve on these worlds.
Because the number of potential sites is so large (20 to 40 billion sun-like stars in our galaxy), this means the odds have to be stacked against the formation of life in a pretty big way for it not to develop elsewhere.
The real question at this point is whether intelligence has long-term survival value from a Darwinian perspective. If being intelligent helps an animal survive periods of scarcity, evolution should favor the development of intelligence given enough time. So, then the question becomes "Does intelligence threaten a species' survival?"
Based on our own experience, the jury is still out on this. We have developed the technology to destroy ourselves, and while we've restrained ourselves so far, it's easy to imagine different scenarios. What would have happened if a dictator like Saddam Hussein had taken control of the former USSR as it was collapsing? Probably something unpleasant. Our society is a lot more fragile than we lead ourselves to believe.
If technological civilizations manage to survive the type of period that we're in, there may be many sites that we can potentially communicate with. But, if self-destruction is the norm, stable technological civilizations may be very rare. Either answer is very valuable. If we learn that there is no intelligent life, we need to understand why, and what implications this has for our own long-term survival.
So, learning whether other technological civilizations exist has important uses beyond answering the quasi-religious question of whether we're alone in the universe. It could inform us about how to avoid the pitfalls that have the potential to hurt us.
oreilly.com: Do you think intelligent life exists elsewhere?
McConnell: I am convinced that intelligent life has developed elsewhere. I can't prove it, but it just makes sense when you consider how many places there are for life to develop.
Humans aren't the only intelligent animals to evolve on earth. Many animals are relatively smart. We often define intelligence based on literacy, so we say that dolphins, primates, and even some species of parrots are dumb, when in fact they are pretty smart. They can remember things. They can communicate. Some of them can categorize objects. They can do many of the things we routinely do, but they can't read a newspaper. And because they can't read and write, they haven't developed technology and cultural communication, which is why our society is so different from other primate societies.
While I think that intelligent life almost certainly exists elsewhere, it may turn out that our type of intelligence is unique. Literacy may be a fantastically unlikely invention.
This is also interesting. It means that if SETI fails, after decades of searching, we will have failed to find intelligent, communicative civilizations (those that have developed the capacity for interstellar communication). This will not necessarily mean that intelligent life, on par with dolphins for example, does not exist elsewhere. That will still be a mystery.
oreilly.com: What role does SETI have in space exploration?
McConnell: Two of the dominant themes in space exploration in the twenty-first century will be the search for extraterrestrial life, and the colonization of our solar system. We will send probes to Mars, Europa, Titan, and other interesting places in our solar system to search for signs of past or present life. We'll send people to Mars, and eventually other places, to continue the search in person. Robots can only do so much.
It's unfortunate that SETI has been dismissed by political leaders, including some of the leaders in our space program, because it's rare that a science program generates such strong public interest.
Whether people like to admit it or not, the public is very interested in the subject of extraterrestrial life. Tapping into this interest could reinvigorate interest in the space program during a time when its funding will be under attack by budget-conscious politicians.
SETI is cheap, dirt cheap, compared to other exploration programs. Yet, the potential payoff is huge. Most people are surprised to learn that SETI receives essentially no public funding. Nearly all the work is funded by private donations, most of them from successful entrepreneurs in the technology industry, like Paul Allen, Nathan Myhrvold, and many others.
SETI can also piggyback onto other space science programs. In the upcoming decade, we will be launching the next generation of space telescopes. These platforms should be outfitted so that they can be used in tandem for SETI research. The Terrestrial Planet Finder, due for launch in 2011, is a good example. This space telescope is designed to filter out the background glare from a star, which will enable it to see the light reflected off of earth-size planets. The same technology can be used to detect pulsed or continuous visible and infrared lasers (one possible method of interstellar communication). A piggyback SETI detector will look at whatever star the primary telescope is looking at, without disrupting the observations of the primary users. If it spots something interesting, it will send a message down to the telescope operators.
Think of this as "SETI Inside." SETI detection equipment can be embedded in space- and ground-based telescopes for a comparatively low cost. This approach will enable all of these instruments to passively search for artificial signals. The incremental costs of doing this are minimal, so it makes sense to do this.
oreilly.com: Why doesn't SETI receive government funding?
McConnell: It's a pretty inexpensive program compared to virtually any government operation. I think a lot of this can be attributed to the "giggle factor." A lot of people have been trained to associate SETI with little green men, cattle mutilation, and uninvited anal probes.
This perception was unjustified, as SETI has its roots in solid scientific thinking. Phillip Morrison, of MIT, coauthored the original paper about radio-based SETI in 1959. Charles Townes, coinventor of the maser (precursor to the laser) and Nobel laureate, coauthored the original paper about laser-based interstellar communication in 1960. So, the field has always been based on real science.
What's interesting is that a majority of Americans think that extraterrestrial life exists (and a sizeable fraction of them think that it has already visited earth). So, the idea of funding a research program designed to detect extraterrestrial life really shouldn't be that controversial. If it were a $10-billion-dollar program, that would be one thing, but you could run an impressive SETI program with a fraction of that amount, or less than a dollar per taxpayer per year.
The lack of funding isn't all bad. Private donors tend to make their decisions faster, and they impose much less arcane conditions on the use of their money. So, being strapped for cash has forced SETI researchers to be more entrepreneurial, and a lot more efficient in the way they use their money, which is why the researchers at U.C. Berkeley built SETI@home instead of buying an array of supercomputers on the taxpayers' dime.
oreilly.com: Do you think this will change?
McConnell: If people make a stink about it, I think so. Politicians react to the polls, and the polls clearly show that a majority of people are interested in the subject of extraterrestrial life.
The main thing that SETI researchers need to do their job is better access to facilities, specifically telescopes. Currently, researchers are forced to borrow time on telescopes normally used for other research, or to devise clever schemes that allow them to piggyback on other projects.
The government could help in a couple of ways: first, by funding the construction of telescopes and related facilities that are reserved for the search for habitable planets and extraterrestrial life, and second, by adding piggyback detection systems ("SETI Inside") to the space telescopes slated for launch in the upcoming decade. Compared to other programs, such as the International Space Station, this can be done cheaply and relatively quickly.
The political climate has changed a lot since SETI funding was killed by Congress. There is tantalizing, though inconclusive, evidence that life existed on Mars. We have since discovered about 50 planets orbiting other stars. And, most importantly, from a politicians point of view, there are nearly three million prospective voters who are running Berkeley's SETI@home software on their personal computers. But, although things have improved, and the overt hostility toward SETI has abated, the program still receives essentially no public funding.
oreilly.com: What could we learn from another civilization?
McConnell: One of the things I discuss in my book is how it is possible to communicate in greater depth than most people imagine. If there is another technological civilization nearby, it will be possible to communicate using equations, images, and algorithms. Computer programs are very useful as the basis for a communication system. I discuss this at length in my book.
So, what could we learn from them? A common assumption is that they would fax over instructions on how to build a fusion reactor, or some other exotic technology. While it's possible they would do this, they might also assume that we can figure physics out on our own. Just as it would be a waste of time for a technological civilization to teach another one how to do basic arithmetic.
The most interesting information will be about things that we cannot observe here on earth. For example, it would be fairly easy for a civilization to send a form of encyclopedia describing the history of evolution on their planet. Once you work out a system for sending images and equations, it is pretty easy to arrange these into a structured message, not unlike a card catalog. It might not be elegant prose, but it will be sufficient for describing what different animals looked like, when they lived, and how they related to the surrounding environment.
While the laws of physics and chemistry are the same here and there, we have no way of predicting the details of how life would develop and evolve on other worlds. Likewise, other civilizations, even if they had mastered everything there was to know about physics and chemistry, would probably be interested to know what types of animals evolved here on earth. That's just one example.
One of the things I discuss at length in my book is the idea of using algorithms (short computer programs) as a foundation for communication. Using programs, you can send what are, in effect, living symbols. This enables the sender to transmit information very efficiently, and also to describe things that cannot be defined with static images or equations.
The variety of ideas and information that can be transmitted this way is essentially infinite, and is only limited by the creativity of the author. What we could learn from this is anybody's guess, but if we do stumble across an information bearing signal it will probably be very interesting.
This assumes that another civilization wants to communicate with its neighbors. It's also very possible that we will detect an unintelligible signal, or a signal that does not carry any useful information (such as a radar signal). In that case, we'll still learn that there is another civilization out there, but we probably will not be able to understand what they are saying. However, if they want to communicate, it will be possible to do so.
oreilly.com: Most language translation programs fail miserably, so how can it be possible to communicate with another civilization?
McConnell: This is a good question. How can we hope to decode an ET message when we have yet to write software that can accurately translate documents from one human language to another?
The reason these programs do poorly is because the documents they are translating are not written in a standardized format. Because of this, the programs have a difficult time determining the meaning of specific words and phrases. For example, does the word "lie" mean "to lie, be dishonest", or does it mean "to lie down". Same word, but two completely different uses.
This problem goes away if the source message is encoded or marked up in a way that guides the recipient in decoding the message. One of the ways you can do this is by creating a numeric language, where each idea is associated with a unique numeric address. This eliminates confusion about the meaning or context of a particular word or symbol in an expression. It increases the author's workload, the process of writing such a message is painstaking, but it makes things much easier on the receiving end.
This general technique of making messages that have embedded clues about how to decode them is known as anticryptography, something I discuss in Anitcryptography: The Next Frontier in Computer Science. What this means is that if we happen to stumble across communication that was not meant for us, we probably won't have any idea of how to decode it. On the other hand, if a civilization is making an attempt to initiate communication with us, they will most likely organize the message in a way that makes it easier to acquire and decode. If that's the case, then we could learn quite a bit from their message. It may take us many years to figure it out, just as it will take us a long time to comprehend the information contained in the human genome. However, once we have captured the raw data, it will be sitting there quietly challenging us to figure out what it means
oreilly.com: How would SETI's success affect society as a whole?
McConnell: I think most people will react with fascination, and probably some fear, if we detect a signal from another civilization. But overall, I think we'll adapt to the new situation, and what was once unusual will become part of our reality. I think it will be easier for us if we know what's in the signal because this will tell us a lot about the intentions of the other civilization.
As for how it would change our society, your guess is as good as mine. We have no idea what information would be in a message, or if we could even comprehend its contents. It's really hard to predict how this would change things in the years that followed a detection event.
Now, if our first contact involves UFOs lobbing bombs or nasty creepy-crawly things onto our major cities, then I guess people will have a negative reaction to the whole thing. But, I don't think that's a likely scenario.
oreilly.com: How can spin-offs from SETI benefit society at large?
McConnell: By experimenting with different ways to send useful information across an interstellar communication channel, we will develop improved techniques for developing and distributing software. Again, the general concept employed in interstellar communication schemes is called anticryptography. The idea is to compose messages that describe themselves, at least in part.
This has many practical uses here on terra firma. One example I use is software for transmitting images across computer networks. It is possible to build a system that eliminates the need for people to download special software to view compressed images. The trick is to interleave the data used to describe a message, and the algorithm (or instructions), for translating that data into a recognizable image. In this scenario, the user's computer automatically learns how to process and display new types of images. It's a shift from the current system, where the user must download special software for each type of media, such as needing QuickTime to display one type of video clip, and RealPlayer to display another.
Even if SETI fails to detect anything, it is a useful metaphor for challenges we face in telecommunication and software development here. By designing systems that describe themselves, we will be able to develop better software, and better services based on that software.
If SETI searches do detect something, software developers will play a crucial role in decoding the contents of the message. Because the software development community is so well networked, almost as soon as we detect something, groups of programmers around the world would begin picking through the data. The raw data from any signal will quickly be made available to anybody with an Internet connection.
So, if we do find ourselves on the receiving end of an ET message, I am betting that the message will be decoded by an independent group of programmers.
oreilly.com: You mention software developers. Why will they play such an important role in SETI?
McConnell: Currently virtually all of the attention in SETI research is focused on detecting a signal, not decoding it. This makes sense since there is no point in trying to decode something that we haven't received yet. However, if and when we do detect a signal, the game will change immediately. There will be an intense effort to decipher the contents of the signal.
Software developers are in an ideal position to tackle the job of parsing and deciphering an ET message. If we detect an information bearing signal, the raw data embedded in the signal will be made widely available, whether the powers that be encourage its distribution or not. Anybody with a computer and a compiler will be able to take a stab at the problem. So there will quickly be a very large number of people trying to figure out what is in the message.
This is one of the reasons I was glad that O'Reilly decided to publish this book, because they have such a following within the open source community. One of my goals in writing this book was to prompt software developers to think about and experiment with this problem. Even if we never detect anything, it is still fun stuff to play around with (a lot more fun than rewriting that marketing report CGI script for your boss).
oreilly.com: You talk a lot about cross-over between ET communication and software development. Can you elaborate on that?
McConnell: The main challenge you need to deal with in ET communication is constructing a message that is easy to decode. One way you can do this is by using numeric symbols and algorithms as the foundation for a system of communication. There are a lot of parallels between this type of system and the way information is stored in DNA.
It turns out that this technique, which can be described as a form of anti-cryptography, has a lot of practical applications. The basic goal in anti-cryptography is to create systems that describe themselves. This technique can be used to create software and systems that can evolve rapidly, and that are easier to distribute and use.
One example I use is the creation of a uniform address space for software components. The way this works is fairly straightforward. Every class library, DLL, etc., is assigned a unique numeric address, much like IP addresses are assigned to devices connected to the Internet. This is a simple thing to do, yet it will solve several problems that plague software developers.
For example, instead of distributing bloated, multi-megabyte installation packages, developers will instead send out much smaller packages that contain only the instructions unique to that program. When a user runs the program, his computer will automatically determine which underlying components are required, and will obtain them from a nearby object server (which is analogous to a domain name server). This approach can be used to shrink distribution packages, eliminate component version conflicts, and to increase effective download speeds.
Another example I mentioned earlier is the idea of a universal media viewer that automatically learns to recognize new media encoding formats.
There are many other situations where this will lead to improvements in the state of the art. One of my goals in writing this book is to prompt developers to think about writing software with the goal of building systems that describe themselves.
If a lot of people start doing this, I am sure we will see some interesting software and fun toys as a result, whether or not the folks at Berkeley ever bag an alien.
oreilly.com: What was the most surprising thing about this project?
McConnell: This is my first book project, so I guess I am surprised that it got published. I had a chance conversation with Tim O'Reilly about the book in 1999. I had an outline and a few chapters, and shared them with Tim. A week later, I had a book contract. The book was published almost by accident.
oreilly.com: Did you have any reservations about writing this book?
McConnell: I had a couple of concerns about this project. It's a controversial subject, and so I had some concerns about the book and the ideas in it being misrepresented. Some people have an automatic tendency to lump SETI in with quasi-religious, "new age" subjects. Many bookstores, for example, initially wanted to shelve the book in their religion and spirituality sections, even though this is a technical book. Some people definitely have a built-in bias about the subject that can be difficult to get past.
It doesn't help that the media is more interested in sensationalism. As a result, SETI gets lumped in with UFOs, and because of this, if you write a book about extraterrestrial life, a lot of people assume you buy into the whole UFO phenomenon. The sensationalism is good and bad. It's good in that it means people are paying attention to the subject. It's bad because there is a lot of bad information floating around out there.
My biggest worry was the quality of the book. SETI touches many different fields of study, from biology to linguistics. While I am an expert on telecommunications, and a pretty good software developer, I am not an expert in these other areas. So, I had to learn a lot to write the rest of the book. This was also my first solo book project. I didn't really know what to expect when the book went out for technical review.
I was pleasantly surprised at how the technical review process went. We had the book reviewed by a really interesting group of people, including some of the leading SETI researchers, well-known science fiction authors, and people from the technology industry, such as Vint Cerf, coinventor of the TCP/IP protocol suite. Overall, they liked the book, and contributed a lot of useful remarks about how to improve it further. That was a nice surprise.
oreilly.com: Speaking of UFOs, what is your opinion about the UFO phenomenon?
McConnell: I am pretty skeptical about it, mainly because the quality of information related to UFO sightings is almost always very poor. That said, I don't think people should discount the idea out of hand. Interstellar travel, though beyond our current capabilities, is physically possible (if you're willing to put up with some very long transit times). Unlikely and impossible are not the same thing. If something is possible, but unlikely, we can't dismiss it arbitrarily.
While I think the accounts you see in the tabloid press are bogus, I don't have a problem with the idea that robot probes could be observing us from orbit. We've sent robot spacecraft to most of the planets in our solar system, and we only learned to fly about 100 years ago. So, it doesn't seem like such a great stretch that somebody might be watching us in this manner.
Look at it from our perspective. If we discovered a habitable planet in a nearby star system, we would probably be very interested in learning more about it. We would almost surely send some kind of probe to collect more information, though it would be expensive, and would take a long time to do. While this might sound like science fiction, we have already developed the propulsion technology needed to do this. Most people don't know this, but we actually built, tested, and almost flew, nuclear-powered aircraft engines. Many credible designs for nuclear rockets have been proposed. The main reason they have not been flown is political, imagine Chernobyl with wings. The technology is there, and if we had to build an interstellar probe, we could do so within our lifetimes.
This incidentally leads to the so-called Fermi Paradox. This is the result of a question asked by the physicist Enrico Fermi, "Where is everybody?" Fermi reasoned that technological civilizations would discover nuclear fission and fusion, and with this technology could travel across interstellar distances at 1 percent to 10 percent of the speed of light. So, if there are other technological civilizations, why hasn't Earth been overrun by them already?
It's an important question. There are as many answers as there are perspectives on the issue.
So, in answer to your question, I don't buy into the UFO phenomenon, especially the accounts of personal contact. But I don't think anybody can say for certain that we haven't been observed in some way. If we were being systematically observed, someone who is going through the expense and effort of doing so is probably going to conceal their presence to avoid spoiling their experiment. So, it winds up being difficult or impossible to prove one way or another.
oreilly.com: What other resources are out there for people who are interested in SETI?
McConnell: There is a lot of useful information on the Internet. The SETI Institute and Berkeley's SETI@home group publish excellent Web sites on the subject. The Planetary Society is another great source of information.
Two books I strongly recommend are "The Biological Universe" by Steven J. Dick, and "Perspectives on Sharing the Universe" by Seth Shostak. Steven Dick's book in particular provides an excellent history of the search for extraterrestrial life.
If you're interested in the technical details about how to build radio or optical SETI detection systems, there are two books that I highly recommend. The Project Cyclops Manual, available from the SETI Institute, provides a detailed introduction to radio-based SETI systems. Published in the 1970s it is a bit dated, and does not reflect advances in digital computing and optical SETI research. The upcoming SETI 2020 manual will provide thorough treatment of both radio and optical SETI systems. This book will be also be available through the SETI Institute.