This defence of Australian internet censorship is more disturbing than the censorship itself

For those who have been living under a rock (i.e. not checking reddit) recently, the government here in Straya is planning to censor our internet access in a clumsy stunt aimed at cornering the “think-of-the-children!” vote. Veteran lefty Clive Hamilton’s defence of the filter, however, is no tabloid hack job: although he couldn’t resist including some choice examples of porn site copywriting, the worst spectre he bothers conjuring is “evidence” that “indicates” some porn-hound boys may develop “perverse attitudes towards girls, such as being disgusted by pubic hair.” Gasp.

Indeed, with such slim pickings for the knee-jerk moralists, it’s hard to figure out exactly who Hamilton is trying to convince. The thrust of his argument seems to be that the selfish libertarians who want the internet to stay free should not be allowed to get in the way of parents who “club together and decide that it is too difficult or untenable for them to protect children by themselves and want their governments to help them”. Exactly why the parents needs the government’s help is never made clear – even if they really are incapable of monitoring their children, in itself a strange assumption, why do they need the government’s help to do so, and why should it involve intruding upon the private lives of all Australians? If they are already clubbing together, why can’t they solve this “problem” on the level of civil society and voluntary participation, leaving the rest of us out of it? Particularly perplexing is this oxymoronic declaration:

[Parents] don’t want to be the household spy and policeman, forever looking over their children’s shoulders or checking to see what they have downloaded on their mobile phones. They want governments to help them.

A word to the wise, Mr. Hamilton: when your opponents are accusing you of authoritarianism, it’s best not to associate the words ‘household spy and policeman’ with ‘government’ in your response. Is he really suggesting that most Australian parents hold a deep desire for governments to take over an aspect of parenting – guiding the child’s media consumption – that requires enormous trust, discretion and communication with the child? And why on earth did he couch this assertion in words that make the police-state overtones of the censorship plan so explicit?

Hamilton’s piece descends almost into self-parody with this brazen admission:

I have deliberately not considered the question of whether it is feasible to effectively filter extreme and violent pornography on the internet.

Why not? Because, of course,

We need a community debate on the question of whether we should do it before we consider the question of whether we can do it because too many internet libertarians and industry spokespeople cover up their refusal to countenance any sort of regulation by insisting that it won’t work.

That’s right: we need to decide that something should be done before asking whether it is possible, in order to prove that the reality-based nay-sayers were only saying it was impossible in the first place because they thought it shouldn’t be done. Where does the (incredibly well-evidenced) fact that it really can’t be done fit into this decision-making rubric? Nowhere. Catch-22.

I can’t even tell who is trying to convince who of what anymore in this bizarre and scary debacle. I just want my usually sane and reasonably free country to give itself a good kick up the arse and close the book on internet censorship for good.

Sydney protest, December 13 (Facebook event)

Great insights about the internet and the Coase theorem in this week’s EconTalk

Ronald Coase is easily my favorite economist, and one of my favorite thinkers in general. Among his many insights was the idea of “transaction costs”, or the cost of an economic transaction which may or may not be higher than the benefit accrued from that transaction.

Coase’s classic example was the firm. Under standard models of perfect competition, he pointed out, firms have no reason to exist. If the free market is always on average more efficient than centralised, command-and-control allocation of tasks and the means of production, it should be cheaper for an employer to contract out for a service rather than hire and organise a group of people to do it. Yet it is clear that a world where every internal task of a firm, from ordering to the mail room to payroll management, had to be accomplished by a tangle of individual contracts would rapidly grind to a halt.

The key, said Coase, are the transaction costs: the cost of organising each individual transaction with a private contractor is greater than the cost of hiring somebody to do it for you, even if hiring may result in a slightly less efficient worker. Coase also pointed out that there is a maximum limit where the decreased efficiency cost begins to exceed the transaction cost, which is why firms still outsource a lot of their activities. It is not efficient, for example, for a private school to run the graphite mine to make its pencils.

In this week’s episode of the EconTalk podcast, internet and organisational guru Clay Shirky describes how the internet is lowering Coasean transaction costs for many exchanges, to the point where ‘products’ that were previously impossible due to the cost of organisation can emerge. The value of this podcast is more in the lucid examples than in the insights, which are not entirely original. I particularly enjoyed Shirky’s description of the process by which Flickr co-ordinates individual activities to produce collections of photography which would have been prohibitively expensive in the pre-internet era. His analysis of Wikipedia also goes deeper than the usual vague references to ‘crowd-sourcing’, and he cleverly uses the unexpected success of voluntary, collaborative activities such as many open source software projects to point out the importance of new behavioural economics insights. Unfortunately, the podcast ends with a strange and poorly explained hope that creative commons-like vehicles can be used to better organise collective action in the real world. Isn’t that where the high transaction costs we are trying to escape were to begin with?

Australia to filter internet access

I made a conscious decision after the last federal election to stop following Australian politics. It took me a little longer than most Aussie children to realise that, like Australian history and Australian film, Australian politics is largely dumb, boring and irrelevant to daily life. So it was a little surprising to read today that our benevolent rulers plan to force ISPs to filter ‘illegal’ material from all internet traffic, and to block pornography and other ‘inappropriate content’ from users who do not specifically request an opt-out. This is idiotic for so many reasons I don’t even know where to start.

Firstly, let’s look at the practical aspects. As anyone who has come into contact with content filtering software at school or in a home knows, there is no way to filter internet traffic with any reasonable accuracy. Rivers of ‘objectionable’ data are going to get through, and lakes of perfectly unobjectionable content will be dammed behind insensate walls. If this ridiculous plan is actually implemented, this will result in incredible pressure on the government to ‘do something!’ about all the nasty stuff still getting through, and to ‘do something!’ about all the good, important stuff still getting blocked, which will drive the government to ‘do something!’ to make the filter more complicated, more expensive and a bigger speed hump for Australia’s already sluggish internet access. To top it off, for anybody who actually wants access to ‘objectionable’ material, there is no shortage of workarounds ranging from simple browser extensions to the old-fashioned methods of bringing smut across national borders. According to the government’s own report, none of the candidate filters work on non-web protocols such as P2P or instant messaging networks. In the worst case scenario, the government will thus begin cracking down on circumvention technologies themselves – and we all know how well that’s worked in the past. In short, any ‘great firewall’ will be both ineffective for its intended purpose, and a hassle to everyone regardless.

Second come the reasons for having such a filter at all. According to the current plan, there will be two blacklists: one covering illegal material, which will be universally applied, and one covering ‘not child-friendly’ material, which will be opt-out. What possible justification can be given for this second filter other than the most blatant, cynical populism? Anybody who wants to filter their children’s internet access can go out and buy an inaccurate, easily circumvented filter to install in their own homes. Why do we need an inaccurate, easily circumvented filter for every ISP? What if parents want more fine-grained control over their filtering? Who takes the blame when parents rely on centralized filters, and something slips though the net? What poor souls at the ISP will be forced to field tens of thousands of tech support calls from confused, angry parents? How is will the opt-out be administrated anyway? Will the government subsidise ISPs for the administration costs involved?

When it comes to the illegal blacklist, the ramifications do not need to be spelled out. I’ll simply say that anybody who thinks the chance of a ’slippery slope’ into Orwellian state censorship of thought is overblown should note that this is exactly what is being explicitly proposed. No slippery slope is required. Sedition is illegal under Australian law; the filter, as described, would universally block access to material critical of the government. Information about other illegal activities, such as drug use or euthanasia, could also be blocked. It would be naive to think the filter will meticulously remove all content of dubious legality; given that it is a purely populist proposition, it would be a counterproductive to introduce any drastically unpopular restrictions. On the fringes, however – and to many people sedition, drug use and euthanasia are decidedly fringe topics -  there is plenty of room for incredible restrictions on free speech which would go largely unchallenged by a majority of the population.

For what it’s worth, should the great firewall ever come to fruition, I pledge to post any and all circumvention methods I hear about to this blog. Let’s hope that it never has to come to that.

ADDED: Check out this blog, this activism site and this amusing parody site.

I’m putting in the poll below purely to try out the new WordPress polls thing. Have fun with it.

Ubiquity, Mozilla’s new command-line tool, is the future of Firefox and the internet

In the days after Firefox 3 was released, I thought something like this:

“Well, that’s about it. They’ve solved all the memory hogging issues, tweaked tabbed browsing to as good as it’s going to get, and added a few little touches like a decent bookmark manager. Sure, the awesome bar is awesome, but there really isn’t anything Mozilla could add to this that would improve browsing without getting in the way. I guess the future for Firefox will be keeping up with the latest standards, and maybe restyling the interface every now and then.”

Turns out I was dead wrong. Ubiquity, although only in a barest of bare-bones alpha release, proves that there is still a vast untapped space of cooler and shinier – not to mention faster and more powerful – ways of interacting with the internet.

Ubiquity is a text command tool not dissimilar from quick launching utilities like Quicksilver or Gnome-do, but with a vastly greater power and integration into the web. Quick launching tools tend to act as intermediaries: they take your command, figure out where you want to go and send you there. Ubiquity is more butler than messenger. You can highlight a section of text and tell Ubiquity to translate it into French, then tell it to email that same section of text to a friend. You can look up a term in Wikipedia and do a word count on the article. Admittedly, these examples are not all that impressive or novel, and the current selection of commands is quite small. However, Ubiquity includes a simple authoring tool and language which lets you create and share custom commands to do pretty much anything. Cool.

After I installed Ubiquity for a test run, the first thing it asked me to do is map a keyboard shortcut to call it into focus. I chose ctrl + spacebar and happily started testing. I spent a few minutes looking through the list of commands (accessible by telling Ubiquity “command-list”), playing around with them and trying to see how many I could use in the course of some normal browsing. I quickly realised two things: firstly, ctrl + spacebar is a very awkward shortcut, and secondly, the need for having a shortcut at all was not entirely clear. If you type a string of text while browsing, most of the time it will just float off into the ether. It’s rare that you would have a text box or some other kind of input selected. Moreover, a lot of Ubiquity commands act upon text you have grabbed with the mouse, which automatically means any text input into the keyboard will go nowhere fast. So why can’t Ubiquity accept text input without calling the command line first? Interacting with Ubiquity doesn’t have the same feel as interacting with, say, the Linux command line, where there is no distinction between the command entry space and the output space. Ubiquity feels more like a utility that you have to call upon than a direct interaction with the browser.

The main flaw with Ubiquity is the same one that has confounded many of the previously mentioned quick launching tools, as well as pretty much any other piece of software that attempts to turn your unbounded input into useful action. There doesn’t seem to be anything particularly unique about Ubiquity in the way it handles natural language commands, which make the eventual universal access dreamed of by its developers seem a little disconnected from the actual software. Without a well-implemented parser or a clear idea of how to subtly force users to make their input machine readable, the ability for Ubiquity to act upon natural language commands will be strictly limited by the foresight and contingency of the commands’ designs. When I first installed Ubiquity, I selected some text and told it to “translate into french”. It took the highlighted text and replaced it with the words “into french”. I tried “translate english into french” and got a similar result. When I read the dropdown menu the third time I typed “translate”, I saw it telling me to enter (text to translate), (to language) and (from language). Typing “translate this english to french” finally got me what I was after, but by this point the original text had long since been “translated” out of existence. Hardly intuitive.

All this having been said, Ubiquity is still a brilliant idea and a groundbreaking piece of software. If Mozilla makes this the centrepiece of Firefox 4 (or whatever it is going to be called), it will probably be their most significant contribution so far to the average user’s experience, not to mention a “into french” against Internet Explorer. Ahem, that’s a “coup de grâce”. They’re not quite there yet.