I've been particularly interested in how the Internet is affecting law and society for several years now. I've seen a lot of debating, litigating, and legislating, but there is little or no consistency in the rhetoric.

In the past, we've had legal frameworks that allowed us to approach these issues with some degree of consistency. Copyright was built on an ethical framework built on certain simple ethical principles: A creator should be given the opportunity to benefit from his work. A creator deserves credit for his work. The market should be grown. A created work must eventually be released back to society, as the goodness of public ownership exceeds the goodness of continued private ownership for the society. These and other simple principles are the foundation of copyright law, even if we have not followed them perfectly.

But communication has changed radically since the precepts in the previous paragraph were first propounded. In all the debating, litigating, and legislating, virtually nobody has examined the ethical foundation of modern communications. Is it any wonder that the resulting legislation and court decisions have been largely garbage?

This essay exists to correct that major oversight. Indeed, the situation has changed, and due to the complicated nature of modern communication technologies, attempting to create an ethical foundation is decidedly non-trivial. This is not to say that the foundations we built the current legal system on have somehow become irrelevant, but how the principles play out in real life is no longer obvious. Rather than starting with a desired conclusion and making sure I end up with them, I examine the situation afresh, hopefully shedding light in corners you didn't even realize existed. In many cases I myself was surprised by the results, particularly by the Death of Expression chapter, when something I thought would be easily salvagable turned out to be a complete loss. The final goal of this essay is to try to construct such an ethical foundation, and from it derive some sort of useful framework for thinking about the ethical problems, leaving the final construction of the legal framework to Congress and the courts.

I do not respect a person, group, or ideology that defines itself solely in opposition to something. Opposition should flow from positive opinion about what should be that happens to contradict somebody else's ideas. To respect my own opinions about communication issues, I feel it is necessary to thus propose a system that describes how we should be doing things, because it's simply too easy to take potshots at an existing system when you have no responsibility to replace it. It is also too easy to paint one's self into a corner when one is simply being critical of things; without a coherent vision, it's easy to end up accidentally contradicting one's self, which I believe has happened many times, even by groups like the EFF.

I run a weblog called iRi, which is descended from a weblog tracking these issues and trying to synthesize all the various stuff that's happened over the last few years into some sort of positive, cohesive whole. It became clear to me quickly that the weblog format was insufficient for the task. There is so much to cover that it simply cannot be done in small chunks, because there is so much context I need to lay down. So instead of trying to lay this out in a long series of weblog posts, I started writing this essay. As you'll see, each chapter builds on the last and there's a high degree of cohesion in this essay; trying to post it in little chunks would never have worked. Now that this essay has been written, I have stopped posting so much stuff about these topics as I feel it is redundant to what I have written here.

I do not ask you to swallow my views unchallenged. Indeed, I encourage dissension; the odds of this essay being perfectly correct are small. But I hope this will stimulate you to think about these issues in new and hopefully profitable ways. There are ideas in here that I have not seen anyone talk about, even in the years of debate I've been watching, and even in the years it has taken me to write this essay, and the years it has been sitting here largely completed. Even a year after the first publish date I think it is still ahead of its time.

This essay alternatively talks pure ethical theory and about legal concepts, both current and future. The connection is that law is always used in this essay as "applied ethics". In the real world, it is not always the case that laws reflect applied ethics; one need not search too hard for laws that seem to be applied corruption. But we will treat the law in this manner, especially as the ethical "action" of the law is often more revealing then the "words" of theory, and transitions will not be labelled, as that would cause too much bloat.

Who This Is For

This essay was written in the United States, about the United States. All references to "this government" and "the people" refer to the United States Government and United States citizens. Conditions in your country may vary... but as I am talking ethical theory here, the contents of this essay still apply.

The first question that needs to be answered is, "Is there any need to re-analyze the ethical situation? Can our current legal/ethical framework handle the challenges posed by modern communication technology?"

In this chapter, I'm going to cruise through history and extract an overview of the development of the concepts that the legal system has developed to handle various issues as they arose. We will re-examine the history of content distribution and the history of copyright principles, and draw some connections between them, both traditional and novel. We will examine how we lost track of the ethics of copyright and got stuck in the trap of believing that ad-hoc, expedient solutions were instead immutable wisdom of the ages. In the end it should be clear that the introduction of the Internet and associated technologies constitute a qualitative change in communication technology that will require extensive further refinements

This is a high-level summary, not an enumeration of the thousands of details added over the years. The intention is to lay the groundwork for a demonstration of the deep flaws in the current system. Because the conventional view freely mixes laws and ethics, this chapter will not go to great lengths to separate them either.

I'm going to assume that the readers of this essay are already familiar with both the justifications for intellectual property and free speech, and the basic historical reasons for the creation of each of those, and do not need Yet Another (Probably Oversimplified Anyhow) Explanation of why Gutenberg's printing press more-or-less caused the creation of the concept of copyright. (Again, this is already long enough without rewriting that yet again.) Therefore, since you know the basics of the conventional view of copyright, we can focus on synthesizing a new and better understanding of communication issues, rather then re-iterating conventional understandings.

Definition of Information

Information is used loosely in this essay to mean anything that can written on some medium and transmitted somehow to another person. Writing, sculpture, music, anything at all. Information can be communicated, which will be more carefully defined later. Yes, this is broad, but there is a rule of thumb: If it can be digitized, it's information.

Note that digitization is very, very powerful. While few people may own the equipment to do it, there is no theoretical difficulty in digitizing sculpture, scent, motions, or many other things people may not normally consider digitizable. Even things like emotions can be digitized; psychiatrists ask their patients to do so all the time ("Describe how anxious you're feeling right now on a scale of 1 to 10."). While something like a written letter may be fully analog, one can generally create some digital representation that will represent the letter satisfactorily, such as scanning the whole letter and sending the image file.

"Communication" is simply the transfer of information from one entity to another. The details of that transfer matter, and it's worthwhile to examine the history of those details.

Historical Overview of Information Transmission

Reduced to a sentence, Gutenberg's printing press's primary effect on information reproduction was to make the production of words relatively cheap. For the first time in history, the effort required to make a copy of a textual work was many times less then the effort required to create the original copy, thus making the production model of "Make an original copy of a book, then print thousands of copies of it quickly for a profit" practical.

Ever since then, technology's primary effect is to lower the cost of various production models with various media until they are practical for an increasing number of people. Every major challenge to intellectual property law has come from this fundamental effect.

Hand Copying

Let's create a graph to look at how various parameters affect the ability to do something profitably, examined over time. (Though not strictly in chronological order as it is often difficult to place a date on when a given technology became truly practical.)

Pre-Gutenberg Practical Information Transmission

You'll note there's only one black-dot, which is hand-copying material. You'll also note that the graph has no axis labels. This is because there are, in reality, a lot of things that could go on those labels, like cost-per-copy (extremely high), time per copy (a lot), number of customers waiting for product, etc., and there are even more as time goes on, so we're just going to pretend that the graph is two dimensional so we can fit it on the paper. I'll talk about some of the noteworthy parameters as we go through.

At this point, there was no such thing as copyright. Copying was an unmitigated good for a society. Numerous documents have disappeared from history because they weren't copied, now existing only as obscure references by other works which were successfully copied.

Gutenberg: Books, Magazines, Newspapers

Gutenberg's invention added another possibility. By lowering cost-per-copy and time-per-copy by orders of magnitude, it became practical to run thousands of copies of a book and sell each of them at a lower price then a single hand copy would cost. A significant time investment was required to set up each run, though, and that became a new constraining factor. With this new ease of replication, the first rumblings of copyright law began, but it was still a very simple domain, so the laws were simple, at least by modern standards.

Another side effect of Gutenberg's invention was the ability to reach an unprecedented number of people with the same message, because of the sheer number of copies that could be cranked out and delivered to people, rather then requiring the users to come to one of the rare copies of the content. This introduces the notion of the scale of communication; throughout history, we have always treated communication reaching many people quite differently from private, 1-to-few communication. Perhaps one of the most important effects was that such technology made it much easier to spread propaganda. Before such easy printing, propaganda required a network of people to verbally communicate it to the targets; printed propaganda, combined with wide-spread literacy, enabled much smaller groups to effectively use propaganda, which has obvious large effects on the fluidity of a society and the intensifying of common discourse.

As the printing press technology improved, people could set up content for the press faster. The lowering of the cost-to-setup enabled the invention of newspapers (and by extensions all periodicals), which are basically cost-effective periodic books. A new practical content distribution solution appeared, and it too affected the law. People wanted to use this new platform for political purposes, but the centralized nature of the printing press made it easy to shut down if a powerful person disliked what the newspaper said. To counter this, our ethical concepts of free speech and the freedom of the press, initially synonymous, were created. In America we even get this guaranteed as part of the first amendment to our constitution; your country may vary. Printing was a major improvement over hand copying, but it is not a perfect information distribution system. The most obvious problem is the need for physical distribution of the printed materials, which was a major part of the cost. The necessities of daily/weekly/monthly distribution to hundreds or thousands of points for periodicals within a subscribing area required a huge infrastructure investment, and non-periodicals needed some infrastructure too, though it wasn't as demanding. There also need to be enough readers (amount of use) to make it economical to print a given newspaper.

Post-Gutenberg Practical Information Transmission

You'll note the hand-copying blob is still there. That's because it was every bit as practical to hand-copy things before the printing press as it was after; that almost nobody chose to do it for book-sized communications means that they felt they had a better choice, but the choice was still there. To this day, hand-copying is still the preferred method of communication in many smaller domains, such as jotting down addresses, phone number, and simple notes. Rarely, if ever, does a truly new form of communication completely displace an older one.

Postal Service

The postal service is not often considered as an important advance in the context of intellectual property, but in communication terms it is the earliest example of a information distribution service that was capable of reliably sending a single copy of something from one single person to another single person. On a technical level, the modern Internet functions much more like a fifty-million-times-faster postal system then the more-often used metaphor of a the telephone system, so study of the postal system can potentially provide insight into the Internet as well.

Postal Service

Whether or not this pre-dates Gutenberg's press mostly depends on what you call the first "postal service". Ancient China had a decent one, as did Rome, but the scale and universality of the modern postal services places them in a different league entirely. The forces of technology have given postal services the ability to serve more people and allow them to send more types of things, culminating in that greatest of postal service triumphs, commercially viable junk mail.

That may sound funny, but it's serious, too. It takes an efficient system to make it worthwhile to simply send out a mailing to "Boxholder", and have any hope of it paying off economically.

We often ignore the media in which normal people can communicate with other normal people on small scales, because the large ones that we are about to look at look so much, well, larger that the postal service seems like it's not worth considering. It's an important advance, though, and has empowered a lot of political action, direct sales, even entire industries that might have otherwise never existed. And it has caused the creation of its own fair share of laws and principles. The postal service is primarily interested in the transportation of objects more then "information" per se, but laws have been developed for strictly communication-based crimes, such as using the postal service to send death threats.

The key things a postal service needs is cheap, reliable transportation and customers... lots and lots of customers. It needs to be economical to process each of these point-to-point transmissions, and this means you need to either make up the cost in scale as traditional postal services do, or charge your customers higher prices as courier services do.

Radio and Television

Radio and television (which at the level I'm covering them are similar enough to treat as two aspects of the same technology) are entirely different beasts.

Television/Radio added

With radio and television, content could be broadcast quickly, even immediately ("live"), to any number of people around the country or the world. The customer had to invest in technology capable of receiving these transmissions, but the general public found radio and television more then compelling enough to invest billions of dollars in. Fast forwarding to today, we find that content with broad appeal can be broadcast profitably, news-type content can be broadcast multiple times per day (entire channels can be dedicated to what is essentially the same hour of content, like CNN), increasing the frequency of transmission to "hours" from "days". "Niche" content can also succeed on a smaller scale, though there is still a relatively high break-even point. The schedules of broadcast content started its intricate dance with the American public as each started to schedule their lives around each other.

Radio and television also broke free of the tyranny of the written word. Radio was one of the first technologies that could handle sound directly (the only competition are records and when you consider that a truly viable technology is a judgement call), and television introduced the even more exciting world of video. It was a long time before this truly strained copyright law, as it was not until the 1980's that the mass-market consumer had any easy, practical means of reproduction of video (via the VCR). Concern about the equivalent of recording from a radio did not exist until the 2000's (the ability to record a digital stream directly from a digital radio), because the mass-market consumer did not have the technology to widely reproduce and distribute a recording with any quality. So we can see that one of the pressures on copyright law is the availability of technology that can produce or reproduce content in a given medium.

Radio and television have their own constraining factors too. The expense necessary to put together even the simplest of professional-quality programs is quite high, which introduces the concept of "cost of entry". In theory anybody could start up a television program or network; in practice, it is vastly more difficult then simply having a printing press print 1000 copies of something. Large transmission towers must be constructed, electromagnetic spectrum must be allocated (extremely limited in television before the advent of UHF), and a large staff must be hired to run this station. Thus, only a limited number of large networks could afford to take full advantage of the medium. This has changed with the wide-scale use of cable, and its corresponding ability to transmit low-quality programs without an expensive transmission tower, allowing "public access" channels (probably only due to Federal mandate, though it's hard to know for certain), but the networks still dominate.

The issue of what we now call "monetization" also appeared, brought clearly to mind by the inability to charge for a physical artifact like the printed word. I am aware of three basic viable models: The advertising-based model, the government tax model used in the UK, and a subscription model where an encrypted signal is broadcast and special decoders must be rented (only feasible relatively recently). This also interacts with the economies of scale; you can't make money advertising if only three people are watching your show, whereas forced taxation allows you to target smaller audiences, as long as you can get government funding. An echo of this problem can be seen on the Internet, except the advertising solution isn't working as well.

Of all of the media discussed in this chapter, I think radio and television have been most deeply affected by the way the industries monetize the medium; one need only compare a day's programming from the BBC or PBS to NBC to see the differences.

Retail Distribution Networks

Retail Distribution added

The retail distribution networks have put content into the hands of "the general public". It has become economically viable to distribute content by having the customer come to some retail outlet that orders many copies of various kinds of content and allows the customer to purchase them and take them home. Much like the postal service, this required the invention of reliable transportation and a large enough customer base to sustain the retail outlet.

And much like the postal service, only more so, this has affected the law by affecting everyone, not just some select group of people with convenient access to content in a copyable form. Perhaps more then anything else, the developments implied by the large scale distribution of content in retail stores, put together with the need for consumer technology such as VCRs to use this content, has brought this down to the level where the decisions made regarding these issues will affect everybody in their day-to-day life. What can I do with this CD? Why can't I send my friend a quote from an electronic book?

The existence of a large-scale distribution network for some kind of content, like sound recordings, tends to imply some sort of standard medium for distributing that content. As more people own players for that medium, the technological pressure to create technology to allow the mass-market consumer to also create content on that medium increases. Thus, a few years after the introduction of the CD-ROM, we get mass-market CD writers. DVD writers arrived even more quickly then CD writers did, relative to the initial introduction of the medium. A large-scale retail distribution method by its very existence tends to create market pressure for the creation of technology that will be capable of allowing the user to, among other things, violate copyright laws.

Telephones

Telephones are much like the postal service. In general, telephones are hardly different at all from speaking face to face, and in general there is little special treatment required to handle them. But I do mention them because of the nuisance issues that the law has had to deal with with regard to telemarketers, scammers, and other people abusing the medium for personal gain. We will find the principles inherent in the laws laid down for telephones useful in some other similar circumstances later, most notably the issues surrounding "forced" communication, such as e-mail spam.

We can go ahead and roll fax machines into this category too; their biggest impact on the law was also a "forced communication" issue, where junk spams could cost the recipient money for toner and paper. As such it is very similar to normal telephones, when it comes to the legal issues.

Final Diagram

Final Diagram of Practical Information Transmission

Our final diagram (cleaned up for convenience) has a number of isolated splotches, and a large number of candidates for what could be considered to be axes:

• cost per copy
• scale of people reached (one, tens, thousands, millions)
• time to produce per copy
• number of paying customers
• initial setup costs
• transportation demands
• economies of scale
• speed of transmission
• availability of (re)production technology to the mass-market consumer
• cost of entry
• method of monetization
• whether reception is "forced"

There are no connections between these splotches, because there are no "in betweens" which are economically viable. Running a television station does nothing to help you run a postal service, because the postal service has "transportation demands" and "initial setup costs" that the pre-existing television station does nothing to defray. Even the two most structurally similar systems, the phone system and the postal system, are very different beasts, and there's no way to use one to help build the other in any significant way. This is important, because the assumption that each of these domains was independent became an unspoken assumption in the law.

What that means in practical terms is that when a legal pronouncement was made about television ("A given company may only own two television stations in a given market"), it had little or no impact on the other communication technologies.

The separation isn't completely perfect, if you try you can come up with some things that affected multiple types of communication at the same time. But even some of the most basic ethical principles were often defined differently for different media; witness the difference between slander and libel, for instance, virtually identical high-level concepts that differ only in whether they occur in spoken or printed word.

I'd also like to point out that there's a lot more on that diagram then I believe most people are considering since they tend to limit themselves to merely mass media like television and the Internet. By adding in some of the other technologies, we'll find that we can actually find a simpler, more general pattern that applies even after the Internet comes into existence, by focusing not on the technology but on the patterns of communication itself.

History and the Internet

Before Internet - Almost Attained Stasis

After The Internet - Stasis Broken

The Internet's effects on law can be understood by looking at this figure (with apologies to those who are printing this out to read it):

Viable Communication Models on the Internet

Here's the "problem": The Internet is extremely powerful technology, and is only getting more powerful. There isn't a single axis of our diagram that hasn't been significantly affected by the Internet. Cost of entry? Believe me, if this was 40 years ago you would not be reading this, because I could never afford to self-publish this physically, and I would simply have never written this. Transportation? It doesn't get any easier then sitting in your own home and accessing the world. Scale? Sometimes people accidentally send email to tens of thousands of people when they intended to only send it to one person; that's how easy large-scale communication is on the Internet.

Remember when I defined information? "Basically, if it can be digitized, it's information"? The Internet is all about the transmission of information. When considered in the context of the rest of the computer revolution, which has digitized everything, from simple words to video to interactive games to, well, everything, the true effect of the ability to transmit information in all of its forms becomes visible: The Internet allows every model to be viable economically, all at once!

Ethical Drift

This also explains the ethical drift that has occurred over the last sixty to eighty years. While hammering out the earliest versions of copyright law, it was critical to create an ethical framework for thinking about the issues. But once the system is built, it is easy to treat the system as the goal and forget about the original ethical foundation that it was built on. Forgetting the true foundations of the system was made even easier by the fact that there weren't very many true challenges to the system; adding one more domain may look exciting at the time but the excitement is contained within that domain. It is easy to see with only a little study of the origins of intellectual property that the mistaking of means for ends is nearly complete in current intellectual property law and trends. Only a small fringe group discusses the ethical issues any longer in terms of responsibilities and the basic goals of the intellectual property legal machinery; the vast majority of the discourse is in terms of the rights of the owners, and protection of rights, and often even the ensuring of profit, which is far removed from the original reasons given for our current system.

Legal Attempts To Deal With The Change

Let's look at the diagram again, but remove the black which represents viable activities.

Activities covered by existing law

You'll note that the law no longer covers the complete domain of activities anymore. This presents some major problems, as people such as MP3.com have found out. As a way of branching out, MP3.com attempted to create a business where they would create high-quality MP3 files of thousands of CDs (at the time, it was technically challenging to create a good MP3 from a CD, as evidenced by the notoriously spotty quality of files downloaded from the original Napster), and when people proved to MP3.com that they owned the CD, MP3.com would give them access to the prepared MP3 files. MP3.com figured this would not be an issue because they made the customer prove that they already possessed a legal copy of the song or album they could obtain. The courts found that MP3.com was distributing music files, which requires a license that MP3.com didn't have. Subsequently, that service has been shut down.

This is a perfect example of what the modus operandi of the legal system has been up to this point: To try to extend all of these little circles outward until they cover all the scenarios that are important at the moment. In this case, the courts decided that distribution of music in this manner was the same as distributing new CDs to the customers. Unfortunately, because each of these sets of laws have been designed with any number of hidden assumptions based on the domain that they were created for, there are unresolvable conflicts everywhere these laws overlap.

For instance, consider "sending digital video over the Internet". In early 2000 in Australia, the part of the government charged with licensing television broadcasters briefly considered trying to require anybody in Australia who wants to transmit streaming video over the Internet to acquire a broadcasters license.

Notice that if one only considers the point of view of those familiar with television, this requirement not only makes perfect sense, it shows a good understanding of how the growth of the Internet could affect television. It is probable that an Internet site in a few more years could broadcast streaming video twenty-four hours a day, seven days a week, at television qualities to television-station sized audiences, effectively gaining all the capabilities of a television station. In only a few more years, this capability will exist for everyone. Obviously, if a country wants to regulate what can appear on television, the country will not want to allow such "stations" to bypass the decency laws of that country just because it isn't a "television station". Thus, it was proposed that the bureau should consider requiring streaming video websites to meet the licensing requirements of a television station. Do not think I am mocking this idea; from the point of view of television regulation, this was an unusually forward-thinking idea, especially for the year 2000.

Fortunately, it was rapidly struck down... because when look at this with anything other then a television-centric viewpoint, it quickly degenerates into absolute absurdity. Does Grandma need a full-fledged broadcasting license just to post streaming video of her grandchildren's birthday party for the father who's on a business trip on the other side of Australia?

And flipping it around, can we completely duck the issue of requiring a broadcasting license by not "streaming" video? In places where two classes of law overlap, like in this example where television law is overlapping with personal law, people will attempt to "dance" from one legal domain to another. So, if the law specifically requires streaming video providers to obtain licenses, a provider may decide to delay the video by an hour, create several files containing an hour of content each, and allow the visitor to view those on the web in a non "streaming" fashion, which violates the spirit of the law, but not the letter. In every way that matters, they are still broadcasting and deriving all the benefits thereof (especially if they also just "happen" to schedule all of their content in advance by an hour so this delay gets cancelled), but they've either ducked the law or forced the creation of unwieldy special-case clauses in law or the policies of some entity like the FCC.

And this is only one small case involving two domains! Imagine all the wonderful conflicts you could create with a bit of creativity, many of which have already occurred, somewhere. I could go on for another twenty or thirty pages proving this point with examples, but that would exhaust us both. If there was only a limited number of these intersections, we could cover them all with special cases as we have in the past, but the combinations and conflicts are nearly limitless, with every year's new technology adding a few more.

Here, you can make your own conflicts. Pick one or two from this list: Video, music, spoken word, deriving works from (such as making custom cuts of movies), text, images, software. Pick one from this list: over the Internet, on demand, traded via peer-to-peer, with some unusual monetization system (such as micropayments). Pick one from this list: to one person, to a select group of persons, to a small number of people in general, to the general public. Odds are very good that unless you deliberately select something that already exists ("Video monetized through advertising to the general public" a.k.a. "Television"), you can find cracks in the law. If you select two from the first list ("Music and spoken word over the Internet to a small number of people in general", which is personal radio station DJ'ing, which is a popular hobby in some subcultures), you're certain to find conflicts. And this isn't even an exhaustive list of possibilities, just what I came up with in a few minutes.

I think it is clear that the current approach cannot succeed. We can not address each problem that arises with a stop-gap solution; each such stop-gap adds two more conflicts to the system. Even if we stuck with it long enough to nail down each problem the hard way, and in this era of multi-year lawsuits that would take a very long time, it would still be messy and inefficient, and would still be subject to radical upheaval by a new technology. We are far better off laying a completely new foundation.

On the one hand, you'd think we all understand what communication is well enough to talk about it meaningfully, since we all do it from a very young age. On the other hand, personal experience shows that people get very easily confused about the communication that occurs in the real world. Most people can't really answer "What happens when you request a web page from a web server?" If you can't even meaningfully answer questions about how things work or what happens, how can you expect to understand the ethics of such actions? And a complementary question, if one must have a post-grad degree in computer science to understand what is going on, how can we expect to hold anybody to whatever ethics may putatively exist?

Since nobody can adequately describe what's going on, the debates almost inevitably degenerate into a flurry of metaphors trying to convince you that whatever they are arguing about is the same as one of the existing domains, and should be treated the same way. The metaphors are all inadequate, though, because as shown earlier, there is a wide variety of activities that do not fit into the old models at all. As a result, the metaphor-based debates also tend to turn into arguments about whether something is more like television or more like a newspaper. There are sufficient differences between things like using a search engine and reading a newspaper to render any metaphor moot, so the answer to the question of which metaphor is appropriate is almost always "Neither."

Before we can meaningfully discuss ethics, we need to establish what we mean by communication, and create a model we can use to understand and discuss various situations with. We'll find that simply the act of clarifying the issues will immediately produce some useful results, before we even try to apply the model to anything, as so frequently happens when fuzzy conceptions are replaced by clear ones. After we're done, we will not have need to resort to metaphors to think and talk about communication ethics; we will deal with the ethical issues directly. As a final bonus, the resultant model is simple enough that anybody can use it without a PhD in computer network topology.

The Model

Example

Let me show you an example of this model applied to one of the most common Internet operations, a search engine query. Let's call the search engine S (for Search engine) and the person querying the engine P (for person). Let's assume P is already on the search engine's home page and is about to push "submit search".

1. P (as sender) opens a connection to S (as receiver) via the Internet (the medium). P sends the search request (the message).
2. S, which exists for the sole purpose of searching the Internet in response to such requests, accepts the connection, receives the request and begins processing it. In the past, the search engine has read a lot of web pages. It puts together the results and creates a new connection to P, who is now the receiver, using the Internet. It sends back the results.

   Technical people will note at this point that the same "network connection" is used, as TCP is both send and receive, so no new "network connection" is ever created. This is true on a technical level, but from this model's point of view, there is a new "connection"; what constitutes a "connection" does not always match the obvious technical behaviors.

On most search engine pages with most browsers, you'll also repeat this step for each graphic on the results page, loading a graphic that's on the page. In this case, the person P is the sender for the first connection, the company running the search engine S is the receiver for the first connection, and the medium is everything in between, starting at P's computer and going all the way to the search engine itself.

This model does not just apply to the Internet and computer-based communication. It applies to all communication. When you buy a newspaper, the newspaper is the medium, and the sender is the publisher. When you watch television, the television is the medium, and the television program station is the sender. When you talk to somebody, the air is the medium and the speech is the message. This is a very general and powerful model for thinking about all forms of communication.

Fundamental Property: Symmetry

We hold these Truths to be self-evident, that all men are created equal... Declaration of Independence

Probably the most important and perhaps surprising result of this analysis is a re-affirmation of something that we all should have understood all along: There is no intrinsic ethical asymmetry in the communication relationship. There is no intrinsic value in being the sender over being the receiver, or being the receiver over being the sender. Indeed, every person and legal fiction (corporation, government, etc.) that ever communicates in any fashion does so as both a sender and a receiver at some point.

I mention this to explicitly contradict the subtle, almost subconscious insinuations by large content owners that they have some sort of ethically advantageous position over consumers that should translate to various special privileges. It can be the case that the sender has something the receiver wants, which puts the sender du jour in a position of power over the receiver, but this is an economic position of power, not an ethical one. It is these relationships that drive the so-called "intellectual property" industries. At any moment the receiver may simply decide not to desire the sender's message, whatever it may be (music, movies, newspapers, etc.), and the economic power the sender has is gone. To the extent this remains a purely monetary concern, the symmetry property is maintained, because after all, the sender desires the receiver's money, too. Capitalism is negotiating the level of desires in such a way that business occurs and money and goods flow.

An example of such a "special privilege": Levies on blank media based on the assumption that the some of the media will be used to illegally copy content, which are paid to the certain large companies who own content. The right to charge what is in every sense a tax is granted to these companies simply because they own content, regardless of whether a given purchaser will actually use if for illegal copying. We consumers don't even subsequently receive the right to place whatever we want on this media based on the fact that we have quite literally already paid for it (at least in the US), which might do something to restore the symmetry; we can still be prosecuted for "piracy". So media companies receive these fees not because it is part of a mutually beneficial bargain, not because it is part of some general mechanism available to all senders, but because they enjoy special, asymmetrical privileges not available to the rest of us.

It is ethically dangerous to promote one person or entities interests over another, for much the same reason that the authors of the Declaration saw fit to put "all Men are created Equal" right at the top. Once stipulated that one entity is superior to another, history shows us time and time again that the leverage is used to gather just a bit more power, and a bit more, and a bit more, and so on and so forth until the inequity is so great the inferior entity rebels in some manner. The authors of the Declaration of Independence were quite familiar with history, and they found this so clear they called it self-evident. If you don't agree that this is bad, I really don't know how to convince you; this is axiomatic.

A Natural Balance

Fundamental Property: Only Humans Communicate

You remember where I first showed the communication model I'm building? See the people on each side of the connection? One must never forget that communication only occurs between people (and corporations in their capacity as people). This may sound like a strange thing for a computer scientist to say, but it is vital to not get distracted by the technology. Computers are null entities in ethical terms. Only how people use them matter.

This is critical because it is so easy to get sidetracked by the technology, but the tech just doesn't matter, except inasmuch as it allows and enables communication among humans. If a computer just randomly downloads something for no human reason (say, some weird transient bug due to a power spike) and it will never be seen by a human, it really doesn't matter. If someone downloads a music file from a fellow college dorm resident and immediately deletes it, it may be technically illegal (in the "against the law" sense), but ethically I'd say that's a null event. If anything occurs that never makes it back to a human being at some point, who really cares?

It is worth pointing out that by and large, current law sees things this way as well. One does not arrest a computer and charge it with a crime (exempting certain cases created in the War on Drugs, which is beside my point here).

The claim that computers can violate the copyright of software by loading into memory from the disk (necessitating a license that permits this act)? As stupid as it sounds. Who cares what a computer does? The only actions that matter are those performed by a human. Media are just tools, they have no ethical standing of their own. In fact, a human never experiences any copy of any communication located on a hard drive. The only copy that matters is the one the human is actually experiencing, which are the actual photons or air vibrations or whatever else used to "play" or "consume" the media.

Only people matter.

An example to illustrate the point: Suppose a hacker breaks into a computer and installs FTP server software on the computer, allowing it to serve illegally copied software and music. Suppose that four days later, the computer owner notices and takes immediate action to shut the FTP server down. During those four days, the computer may have served out thousands of copyright violations. Ethically, can we hold the computer, and by extension, the computer owner responsible? No! We should hold the hacker responsible, not the computer. Only humans communicate, and the computer owner was not even aware of the offending communication, and took no actions to enable it. His or her computer was being used in communication as a medium by completely different senders and receivers. Ethically, the owner was an innocent bystander to the software piracy.

Note this scenario happens daily, and to the best of my knowledge nobody has ever been prosecuted for being hacked and having an FTP server run on their machine. This could change at any time...

The Key to Robustness: Follow The Effects

It turns out this fundamental property is the key to robustness. If we had to pay attention to the actions of every computer and router and device between the sender and the receiver, we'd never be able to sort out the situation with any degree of confidence. The modern Internet is quite complicated, and some innovative and complicated ways of communicating have been developed. As we try to apply older models to these issues, the complexity explodes and we are left unable to determine anything useful about the situation.

But we don't need to worry about the actions of every router and program between the receiver and the sender. All we need to worry about are the results, and who is on the chain of responsibility. The technology is unimportant.

Fundamental Property: Everything Is Digital

Everything is digital. There is no analog to speak of; analog is an artifact of technology, with little to no discernable advantages over digital, except for generally requiring somewhat less sophisticated technology to create and use.

Review

To review:

• Communication can be modelled as a set of connections along which messages flow from a sender to a receiver on some medium. A given interaction may require several such connections to model; even a simple web request requires two connections (one from the person requesting the page, one to return the page to that person).
• Several elaborations on the model exist in the real world, which can complicate the model primarily by adding more entities to the chain of responsibility, who have the opportunity to fiddle with the message(s).
• We take as an axiom that communication is symmetric. There is no intrinsic ethical value in being the sender or receiver, or in possessing some message that somebody else wants.
• There is no analog, only digital, and things that aren't digital quite yet, but could be if enough people cared.
• We observe that only humans communicate. We also note that we either must accept this as true, or give up any hope of having a rational system of ethics that can actually claim something is ethical or unethical.

Now that we have examined the history of communication, laid down our goals in building an ethical system, and created a model we can use to model actions in the real world and communicate about their effects, we are finally ready to begin applying what we've built to real-world situations.

Elaborations on the Model

Chain of Responsibility

There are many elaborations on this basic model:

• an entity between the sender and receiver catches the message and does something with it. The entity could cache it so future requests are faster, log that a communication occurred, change the message, completely block the message, or any number of other things.
• the communication medium itself may affect or manipulate the message somehow, either according to someone's intent or by accident (transmission flaw), especially as "the medium" can include computers and other such things.

There are any number of ways to expand on those basic elaborations of intercepting the message or manipulating the medium. An exhaustive list of such things would take a long time, but fortunately there is a relatively straightforward way of characterizing all of them. Each possible elaboration consists of some other entity inserting themselves into the connection somehow and manipulating the message between receiver and sender. For every communication made, we can catalog these entities and form a chain of responsibility.

chain of responsibility

The listing of all entities responsible for manipulating a message in some form other then pure delivery.

Consider a newspaper story containing a quote from some source and some commentary by a reporter. By giving the quote to the reporter acting on behalf of the newspaper, the source is trying to communicate with the readers of the newspaper story. In the communication from the source to the reader, the newspaper intervenes, probably edits the quote, and adds the rest of the article around it. In this case, the newspaper is on the chain of responsibility for this communication. This matches our intuition, namely that the newspaper could potentially distort or destroy the message as it pleases, and it has responsibilities that we commonly refer to as "journalistic ethics", which among other things means that newspaper shouldn't distort the message.

Being "responsible" for a message means that you are able to affect the message somehow, and are thereby at least partially responsible for the final outcome of the communication. As I write this, millions of conversations are occurring on a telephone. I am not on the chain of responsibility for any of those communications because I have no (practical) ability to affect any of those communications. On the other hand, if someone calls my house and leaves a message with me for my wife, I can affect whether that message is transmitted accurately, or indeed at all. Alas, too frequently I forget to deliver it. Thus, I am on the chain of responsibility for that message, since I have demonstrated the capability of destroying the message before it got to the recipient.

Anybody who can affect the message is therefore on the chain of responsibility even if they have no technical presence on the medium. The biggest, and perhaps only, example of this is a government, which may choose to set rules about all messages that affect all message and thereby have a degree of responsibility for all messages. For instance, the government makes rules about "libel" and "slander", and the government has the ultimate responsibility of enforcing them. Since a government is capable of censoring a message, they are technically on every message's chain of responsibility as a result, though the impact is so diffuse that usually as a practical matter it's not worth worrying about.

The chain of responsibility should consist only of people, corporations in their capacity as people, and "governments" considered as people. Any time it seems like a machine or process is on the chain of responsibility, it is really the person responsible for that machine or process who is on the chain. The reason for this is that a machine or device can not be "responsible" in the ethical sense for anything, since they are not people. Sometimes it is not obvious who that person is, and once again it can be a judgement call exactly who is responsible.

For example, consider a browser cache, which stores content from web servers on behalf of the browser user, so they don't need to reload it every time they wish to view it. Clearly "the browser cache" is on the chain of responsibility of a standard web page retrieval communication, because if it works incorrectly or has stale content, it can prevent the user from receiving the correct message. But since "the browser cache" is (part of) a program, and programs aren't allowed on the chain, who is responsible for the browser cache? The obvious answer is the browser manufacturer, but as long as the cache is implemented correctly, that is not necessarily the right answer. Consider a lawsuit against someone who took content from a web browser's cache and then illegally distributed it. Can the browser maker be said to be involved with this? One could make a case for it... perhaps the cache manufacturer should have encrypted the cached content better so the user couldn't just take it out of the cache. On the other hand, that's not necessarily a very convincing argument because clearly the person who is illegally distributing the content is responsible, and the fact that they got it from the cache merely incidental, as they could have just as easily gotten it directly from the website.

Presence on the chain is not a binary off/on thing, because there are different levels of responsibility. Sometimes it's not possible to strictly determine whether an entity is "on" or "off" the chain. As usual in the real world, there are grays, and thus there is room for legitimate disagreement about how responsible a given entity is in some situations. The most useful question to ask is "How much influence can the person have on the message?" Someone who can silently manipulate the message to say anything they please obviously has more responsibility then someone who can merely block one image from loading on a web page, and can't hide their responsibility. In the caching example, as long as the cache is correctly and honestly implemented, the browser manufacturer has no effective ability to control the messages I see.

Another example: We typically believe that the phone companies or Internet search providers should not examine or modify the messages we use their equipment to send, but just accurately transmit them. In other words, we expect the ISPs and phone companies to stay off the chain of responsibility by refusing to affect the message, despite the fact that physically, our messages travel via their equipment and they could fiddle with it if they chose. This is also a good example of a time when literal physical reality doesn't perfectly match our conception of ethics: If the phone company simply relayed our message, we do not hold them ethically responsible for the contents of the message, event though in a technical sense they are. In existing law we refer to this as "common carriers", entities that simply carry communication and are not responsible for the message, contingent on their not affecting the message.

Time

Intention vs. Literal Speech

If you read "Hlelo, how are you?", you still understand what I mean despite the typo. (I'm sure you can find some more typos in this work elsewhere for more examples.) But sometimes it isn't so easy to tell what the original message was supposed to be, if the typo or corruption is bad enough. Or sometimes the communicator can't or doesn't say what they mean, or it may not be possible to directly say what they mean in a given medium.

It is always impossible for a receiver to be completely sure they truly understand what the sender was trying to communicate with their message. Rather then opening the topic of whether we can get the "true content" from the message itself, which itself has many large, heavily philosophical books written about it, we will dedicate ourselves to the much simpler task of just trying to make sure that the message itself, the sequence of bits over time, is adequately transmitted from sender to receiver, because that's all we can do.

We will see this come into play later, as we try to determine whether something "really" changed the content of a message or not.

In the United States, we have the First Amendment of the Constitution that guarantees us certain things.

Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.

This is less of a concern to some countries, such as China. However, modern communication capabilities can affect free speech in a lot of ways, both enhancing and diminishing, depending on how it is used. No matter how you look at it, freedom of speech will be affected in every country.

What's The Difference?

Censorship and free speech are often seen as being two aspects of the same thing, with censorship often defined as "the suppression of free speech". Perhaps there is nothing wrong with this definition, but for my purposes, I find I need better ones, perhaps something a little less circular. My definitions have no particular force, of course, but when grappling with problems, one must often clearly define things before one can even begin discussing the problem, let alone solving it. Thus, I will establish my own personal definitions as the analysis I want to do is not possible with a fuzzy conception of what "free speech" is.

Free Speech

It's typically bad essay form to start a section with a dictionary definition, but since I want to contrast my definition with the conventional dictionary definition, it's hard to start with anything else. Free speech is defined by dictionary.com as

free speech

The right to express any opinion in public without censorship or restraint by the government.

This definition misses some critical aspects of our common usage of the term. For instance, free speech is of no value if nobody is allowed to listen to the speech; people in solitary confinement have perfectly free speech, but that does not mean that we would have considered it an acceptable solution to lock up Martin Luther King Jr. in solitary confinement and let him preach what he may; along with the obvious unjust imprisionment we would consider this to be an obvious example of trampling on free speech. We should also consider the right to free speech as the right to listen to anybody we choose (subject to possible exceptions later), thus

free speech

The right to express any opinion in public without censorship or restraint by the government, and the corresponding right to experience anybody's expressions in public without censorship or restraint by government.

I use "experience" here as a general verb: One listens to a speech, watches a movie, reads a book or webpage, etc.

Since I don't want to define free speech in terms of censorship, lets remove that and put in its place what people are really afraid of.

free speech

The right to express any opinion in public, and the corresponding right to experience anybody's expressions in public, without being pressured, denied access, arrested, or otherwise punished by the government.

This definition really only applies to people in a government-controlled territory, like a public park. If one looks around at all of the various ways of expressing ourselves, we find that the government does not own very many of them. In common usage of the term "free speech", we expect "free speech" to allow us to say that a corporation "sucks", express our opinions about pop music stars, and review movies, without the non-governmental entities we are talking about, or that own the means of expression, being able to suppress our speech merely because they don't like it.

Considering both the target of the speech and the publisher of the speech is necessary. Suppose I use an Earthlink-hosted web page to criticise a Sony-released movie. If Earthlink can suppress my speech for any reason they please (on the theory that they own the server and the bandwidth), and have no legal or ethical motivation to not suppress the speech, then in theory, all Sony would have to do is convince Earthlink it is in their best interest to remove my site. The easiest way to do that is simply cut Earthlink a check exceeding the value to Earthlink of continuing to host my page, which is a trivial amount of money to Sony. In the absence of any other considerations, most people would consider this a violation of my right to "free speech", even though there may be nothing actually illegal in this scenario. So if we allow the owner of the means of expression to shut down our speech for any reason they see fit, it's only a short economic step to allow the target of the expression to have undue influence, especially an age where the gap between one person's resources and one corporation's resources continues to widen.

Hence the legal concept of a common carrier, both obligated to carry speech regardless of content and legally protected from the content of that speech. The "safe harbor" provisions in the DMCA, which further clarified this in the case of online message transmission systems, is actually a good part of the DMCA often overlooked by people who read too much Slashdot and think all of the DMCA is bad. The temptation to hold companies like Earthlink responsible for the content of their customers arises periodically, but it's important to resist this, because there's almost no way to not abuse the corresponding power to edit their customer's content.

I also change "opinion" to expression, to better fit the context of this definition, and let's call this "the right to free speech":

the right to free speech

The right to express any expression in public, and the corresponding right to experience anybody's expressions in public, without being pressured, denied access, arrested, or otherwise punished by anyone.

There are standard exceptions to free speech, for instance "libel", "slander", "threats", and "community standards." In my opinion, these are not deeply affected by the Internet era, with the exception of what the definition of a "community" is. I want to leave that for later. Thus, my final definition is

Censorship

For structural symmetry with the Free Speech section, let's go ahead and start with the dictionary definition. A drumroll please...:

Censorship

Censorship is the act of censoring.

OK, that was particularly useless. Curse structural parallelism!

The best way to understand my definition of censoring is to consider the stereotypical example of military censorship. During World War II, when Allied soldiers wrote home from the front, all correspondence going home was run through [human] censors to remove any references that might allow someone to place where that soldier was, what that soldier was armed with, etc. The theory was that if that information was removed, it couldn't end up in the hands of the enemy, which could be detrimental to the war effort. The soldier (sender) sent the message home (receiver) via the postal service as a letter (medium). The government censors intercepted that message and modified it before sending it on. If the censor so chose, they could even completely intercept the letter and prevent anything from reaching home.

This leads me naturally to my basic definition of censorship:

Censorship

Censorship is the act of changing a message, including the change of deletion (complete elimination of the message), between the sender and the receiver.

Censorship is not always evil; few would argue that when practiced responsibly, military censorship as described above is truly ethically wrong. Censorship is a tool like anything else, it can be used to accomplish good or evil. But like war, censorship must be used sparingly, and only when truly necessary.

There is one last thing that we must take into account, and that is the middleman. Newspapers often receive a press release, but they may process, digest, and editorialize on the basis of that press release, not simply run the press release directly. The Internet is granting astonishing new capabilities to the middlemen, in addition to making the older ways of pre-processing information even easier, and we should not label those all as censorship.

Fortunately, there is a simple criterion we can apply. Do both the sender and the receiver agree to use this information middleman? If so, then no censorship is occurring. This seems intuitive; newspapers aren't really censoring, they're just being newspapers.

You could look at this as not being censorship only as long as the middlemen are being truthful about what sort of information manipulation they are performing. You could equally well say that it is impossible to characterize how a message is being manipulated because a message is such a complicated thing once you take context into account. Basically, since this is simply a side-issue that won't gain us anything, so we leave it to the sender, receiver, and middleman to defend their best interests. It takes the agreement of all three to function, which can be removed at any time, so there is always an out.

For example, many news sites syndicate headlines and allow anybody to display them, including mine. If a news site runs two articles, one for some position and one against, and some syndication user only runs one of the stories, you might claim that distorts the meaning of the original articles taken together. Perhaps this is true, but if the original news site was worried about this occurring, perhaps those stories should not have been syndicated, or perhaps they should have been bound more tightly together, or perhaps this isn't really a distortion. Syndication implies that messages will exist in widely varying contexts.

Like anything else, there is some flex room here. The really important point is to agree that the criterion is basically correct. We can argue about the exact limits later.

So, my final definition:

censorship

Censorship is the act of changing a message, including the act of deletion, between the sender and the receiver, without the sender's and receiver's consent and knowledge.

In terms of the communication model, censorship occurs when somebody interrupts or interferes with the medium such that a message is tampered with while traveling from the sender to the receiver.

The Difference

Going back to the original communication model I outlined earlier, the critical difference between the two definitions becomes clear. Free speech is defined in terms of the endpoints, in terms of the rights of the senders and receivers. Censorship is defined in terms of control over the medium.

The methods of suppressing free speech and the methods of censoring are very different. Suppression of free speech tends to occur through political or legal means. Someone is thrown in jail for criticizing the government, and the police exert their power to remove the controversial content from the Internet. On the receiver's side, consider China, which is an entire country who's government has decided that there are publicly available sites on the Internet that will simply not be available to anybody in that country, such as the Wall Street Journal. Suppressing free speech does not really require a high level of technology, just a high level of vigilance, which all law enforcement requires anyhow.

Combating Censorship and Free Speech Suppression

The ways of combating suppression of free speech and censorship must also differ. Censorship is primarily technological, and thus technological answers may be found to prevent censorship, though making it politically or legally unacceptable can work. Suppression of free speech, on the other hand, is primarily political and legal, and in order to truly win the battle for free speech, political and legal power will need to be brought to bear.

These definitions are crafted to fit into the modern model of communication I am using, and I have defined them precisely enough that hopefully we can recognize it when we see it, because technology-based censorship can take some truly surprising forms, which we'll see as we go.

Summary

• Free speech is the right to express any expression in public, and the corresponding right to experience anybody's expressions in public, without being pressured, denied access, arrested, or otherwise punished by anyone, subject to somewhat fuzzy, but fairly well-understood exceptions.
• Censorship is the act of changing a message, including the act of deletion, between the sender and the receiver, without the sender's and receiver's consent and knowledge.

What is a Patent?

It is well beyond the scope of this essay to go in depth into the history of patents, which itself can easily fill a book. If you are not familiar with patents, I recommend doing a Google search on "Patents" and "Patent history".

The incredibly brief summary is that patents extend protection to someone who discovers a process that does something, such as a process for manufacturing aspirin. For the period of the patent protection, nobody is allowed to use that method unless they come to some sort of agreement with the patent holder. After that, the patent becomes public knowlege and anybody can use it without penalty. It is a limited monopoly over a process granted by the government in the interests of both furthering progress by giving people economic incentives to do research (by allowing them to benefit from that research), and to benefit society by making sure the knowlege makes it back into the public domain (all patent filings are fully public).

In particular, I'd like to point out that initially, the idea that software could be patented was rejected, as there was no material process that could be patented. A pity this was not sustained, as we would not have the patent problem today.

What is Software?

(Even if you are computer expert, you may want to skim this.)

What Is Software?

When most people think software, they think things like MS Word and Mozilla Firefox. This are pieces of software, yes, but they are not what "software" is anymore then Shakespeare's "Romeo and Juliet" is "English". This comparison is almost exactly analogous to the situation in software.

At the most basic level, software is a sequence of numbers, just as written English is a series of glyphs and spaces. These numbers look something like this:

• 2558595091
• 1073777428
• 1382742636

Software Is Just A Lot Of Instructions

Software truly is a recipe, it truly is a list of instructions. And just like recipes in a cookbook are written in English (or even a specific English-based jargon language), these instructions must be written in some human-created language. It's a limited language. a mathematical language, but a language nonetheless, capable of expressing how to do everything a computer ever does. It is the language of problem solving. It extremely precisely answers questions such as "How do I move the mouse cursor?", "Will the asteroid hit earth?", and "How do I determine the square root of 7?"

Any given CPU understands a special language, called the machine language for that CPU. When a computer programmer speaks in this language, and tells the computer "add 2 and 2 and put it in register 4" (a register is a place to store a number), the computer can 'understand' and do it. Machine language is the compromise language between computers and humans, the meeting point between flesh minds and silicon machines. Like any language, there are things that are easy to say and things that are impossible to communicate. (English, for instance, does not communicate mathematical concepts very well, that's why mathematicians have their own mathematical language that looks like this: )

Machine language's most notable characteristic is that it never says anything implicitly, it is always explicit. For computer scientists, where slight nuances in meaning can mean the difference between a calculation taking years, being completed in minutes, or not being completed at all, let alone correctly, this is an important characteristic of the language, because it means there are never any "devils in the details" that can cause unanticipated problems. (Think about how often you have given instructions to someone, only to have them misinterpret them because they weren't precise enough or were misinterpreted by the other person. Imagine the difficulties that could be caused if that literally occurred several billion times per second, and you start to understand why programmers like the idea of a language that is totally explicit.)

Real assembly code to print "Hello world!" on a Intel-based Linux machine. The three letters following "%" are names of the "bins" I referred to, which are given names rather then numbers. "printf" is a function used to do the actual printing, which itself would be represented by many pages of code like this. All this fragment actually does is set up the processor so the printf routine can be called, it doesn't actually print anything itself. printf itself still does not directly print things out to the screen; it would itself call other "lower level" functions to do that! The "l"s on the end of the command name are an abbreviation for "long", which is a type of number. Take off the "l"s mentally and you can almost make sense of some of this; the computer is copying numbers and subtracting offsets from those numbers to make room for printf to run.

Humans almost always represent machine language in a closely related language called "assembly language", which is slightly closer to human languages. Subtracting 250 from 500 and putting the result in bin 1 might be represented as SUB 500, 250, 1. Assembly language traditions tend to be old and come from an age when the machines were much less friendly, so that subtraction operation is likely to look more like SUB I #500, [#250], {%1}; the wierd symbols may look intimidating but like the exclamation mark and colon in English, once you learn what they mean they do not look so strange. Assembly language can represent a more-or-less one-for-one translation from machine language, and is significantly more convenient then looking at the raw numbers.

Non-Assembly Languages

The original C code to the assembly program in the previous figure.

In reality, assembly language is too low level for normal usage by humans. For example, the code in the assembly language figure only sets up a single function call; that's a lot of typing to do something very, very simple; any non-trivial program will have thousands or even millions of function calls, so a function call needs to be as simple as possible because every keystroke will be repeated that many times. Fortunately, we can create our own languages that are simpler to use for humans, and convert them into machine code. Some famous languages of this type are C and C++, Java, Javascript, Lisp, and Fortran, but there are thousands of others. I've created two of my own for graphics processing and audio processing simply to facilitate finishing an in-class assignment. We weren't required to write a language; it was simply easier then the alternative. It is quite easy to write a language for a specific purpose for a person skilled in the art; I only mention the big ones because I expect you might recognize them but I would imagine that not a day goes by without some new little language being created somewhere. Computers don't understand these languages directly, but we can tell them how to translate these other made up languages into machine code that the computer can understand. (The precise mechanics involved in this process are absolutely fascinating, and ranks as one of the great engineering triumphs of the 20th century... and are way too complex to cover here.)

Since even dealing with assembly code gets old very quickly, we taught the computers how to take care of this for us. In C, I can write this:

area = width * height;

and the C compiler will automatically create the code for doing the multiplication and storing of the value, without the code writing needing to muck about with the exact memory location those things go into, or the exact bin the computation is performed in. And more importantly, the code writer doesn't need to think about these things, leaving their brains free to think about more important, high-level things that only humans are good at doing.

By creating these languages for our own use, which we call higher level languages (because they allow us humans to think at a higher level of abstraction, which we are more comfortable with), we allow ourselves to use words and other symbols where the computer only understands numbers. But before a modern computer can obey our command, the compiler program must convert "area = width * height" back into the machine code numbers that the computer understands, or it will be unable to execute them.

This is some of the machine language code from the Hello World program.

I derived the assembly language in the assembly language figure from the C program in the C language figure above. The C compiler I use works by converting the C code into assembly language, then the assembly language into machine code; I got the assembly language by asking the compiler to stop there. Most of that is just boilerplate; the important line is printf("Hello world!\n");. (The \n tells the computer to go to the next line after it prints Hello world!.) Almost all of the assembly code in the assembly language figure is the translation of that line, pretty much everything except the last three lines, which are the translations of return 0; and is a necessity imposed by the operating system, which always expects a program to return a number back to it. But you can see how it's still easier to read the C then the assembler. Some languages like Python are even simpler, where the code for the equivalent Python program would simply be print "Hello world!". Finally, I let the compiler generate the machine language from the assembly code, a portion of which is shown immediately above.

Brief Review

To briefly review, we've discussed three separate things up to this point:

• Machine code, the actual code the computer executes. This is a series of numbers. The computer is not capable of executing anything else directly.
• Assembly code, which is a nearly one-to-one transliteration of the machine code into something that a highly trained human can relatively easily read. This is generally the "meeting point" between human minds and the computer, although there is nothing impossible about reading machine code dire