ILaw – 2003 June 20 – Larry on code – updated as it goes along

(entry last updated: 2003-06-30 15:34:20)

A walk with Larry through the key concepts of Code and Other Laws of Cyberspace (warning, I’m new to trying this, and I’m unable to avoid injecting myself into this, so don’t expect a transcript.) (Donna’s notes)

A start on the Viadhynathan discussions of the ideology of anarchy, governed by the mind, embracing freedom – the internet as “unregulable”

A dichotomy – if, in fact, the internet is unregulable, why do we need an EFF, anyway? Because, in fact, the internet can be regulated, but you need to look at a new mechanism of control – Larry points to the list from Code; law, markets, norms (society), and architecture. Or, more specifically, implementation of technologies – the motives that underlie how we choose to deploy ways of doing things.

Smoking as a demonstration of the ways that governments employ mixed modes of control to change behavior – taxes, advertising to change mores, regulate packaging, etc.

To demonstrate that architecture is not uncommon, Larry goes into architecture as a regulator chosen by governments or their actors (my architecture lecture – 4 MB pdf). Baron Haussmann, Robert Moses (pictures in my slides), Americans with Disabilities Act

(Side note: Jonathan is "Z", at least when Larry’s picking on him)

Subtleties – the modalities may reinforce or conflict, but governments (and others) will employ these modalities to achieve their policy goals.

A questioner goes after Larry’s claim that “God” is the enforcer of architecture – no policeman is needed to maintain the modality of control that architecture represents. Granted, it can be challenged, and eventually changed, but it does make it hard to discuss, particularly if it’s subtle.

A questioner asks why the law seems to have primacy in Larry’s talk. Larry answers that discussion of modalities of control is somehow more natural in the legal domain, provided within a legitimated framework – e.g., democracy. It is, thus, more legitimated.

(Forgot to mention, Derek came up for at least the day)

Larry on libertarians – in the end, I think it reduces to unwillingness to see

So, getting back to regulability, the internet is, of course, regulable, because it can be (and in fact has been) architected. The first point – architecture has consequences. For example, on the internet, TCP/IP doesn’t let you know who sent a message – there is no identity to IP addresses. You can’t know what’s in a packed (well, you couldn’t originally). You can’t know the location of either the source or the recipient.

This arthitecture of anonymity led to the libertarian argument that the internet is not regulable, etc. Life pre-cookies, in other words. (Wow – Larry stuck his neck out and called http stateless.)

But, the conflict between what business wants to use the net for, what government wants to use the net for and what the designers put into the internet, leads to a tension that (unsurprisingly) motivates engineers to start acting to modify the architecture.

Larry’s point 1: the error of the libertarism argument is “is-sim” – the internet won’t change – and, as we’ve seen, architecture can be changed, because implementation of technology is mutable – that’s what engineers do. Government and commerce have reasons to motivate change – e.g., cookies (to get around the statelessness of http – Netscape to ease the life of commerce servers), packets sniffers (to see the content of traffic – Larry’s Morpheus server story), and IP mapping (Jonathan’s research into blocking).

(Hypothesis: Larry throws Jonathan’s name into the lectures as often as possible, because he’s used to Jonathan tuning out and he wants to make him jump – a classic lecturer’s trick. Empirical evidence shows, so far anyway, that it still works on Jonathan, too!)

These three changes, among many, many more, take away more and more of the anonymity of the internet – and the consequences of the architecture have now changed the internet into a more regulable space – so much for “is-ism.” And, therefore, we better start paying much more attention to the implications of changes in the architecture.

A question on cookies and the Hamadi case: rephrased as “is the cookie a violation of rules or laws?” And the answer is, generally, no. But, the unintended consequences of these changes may be unattractive (note to self – this is a key set of ideas around the ESD notions of engineering systems, something that I need to expand upon). DoubleClick has learned from the questions that their use of cookies raised, and the FTC & norms/reputation (as elements of other modalities of control) have influenced these changes.

Larry’s point 2: the modalities of control interact. With e-mail spam as an example – libertarians believe in free speech, so keep the law out of email. In the early days of the internet, the small homogeneous internet community flamed the first advertising e-mails. But, with the influx of the hoi-polloi into the internet, the norm against e-mail ads vanishes and spam emerges. Responses include white hat vigilantes (real-time black hole lists, other tech fixes) but all their actions really do is make things worse. And, it looks like censorship, an ideological violation – and it works to get you off (John Gilmore story – sp?)

What if a law could reduce the market value of spam? Wouldn’t that allow us to avoid using the arhitecture? Say, Larry’s spam labeling act? Although we aren’t going to talk about that for a while, it appears.

Larry’s summation:

  1. Code is law

  2. Code is plastic

  3. No law can beget bad code

  4. Good law can avoid bad code (maybe)


  1. The US is moving away from respect for public law, so why is the primacy of law asserted? Larry: I am a pessimist; you are right, the law is losing out and perverting many important and fundamental values. (Unspoken: the law is undermining itself because of this failure to reconcile what law does and what consequences emerge)

  2. An ex-network admin from a university describes caving in to Orbs, whose egos got away from the notions of cooperation – the damage of the vigilante groups. Spews (sp?) vs. Paul Vixie’s group are contrasted – leading to a notion that at least governments have to justify what they do, while vigilantes don’t have to.

  3. Another example of (il)legitimacy – the influence of US politics upon the global internet is discussed. Larry points out that the internet is the most efficient mechanism to export the US vision of free speech. Should we homogenize the network protocols, or should we allow jurisdictions to dictate the architecture? An open question.

  4. The expansion of the vigilantiism to the level of international jurisdiction issues. This leads us into a further discussion of jurisdiction – why isn’t it just like the Atlantic ocean (doesn’t the US Law of the Sea still obtain?) Larry points out that location is tied to many of the notions of jurisdiction, and thus many of the natural analogies don’t work when you consider the internet, because the metaphors of location don’t really work seamlessly with these problems.