A colleague has co-written a book that dives into what I wrote about – it’s called Apprenticeship Patterns.  I think I shall have to give it another read through, I feel like I might want to write more about this in the future.

I should start by recognising that laws, however well intended or conceived, will always leave open the possibility of misuse or abuse – quite aside from anything else, progress is naturally disruptive to prescriptive legislation (we need only to look as far as copyright law’s complete inability to deal appropriately with file sharing to see this in action).  This is further complicated by the fact that we characterise a suit as a SLAPP by the intention of the plaintiff (any lawyer will tell you how hard it is to determine intention) – and, of course, SLAPPs aren’t even about winning.  This means that in combating them, we want to fix laws which become overly restrictive or illiberal (common law surfaces such things quickly), reduce the cost of defending against such suits and penalise those who engage in them – the ultimate aim being to discourage them from being filed at all.

Reducing costs for the defence

One of the problems I identified in my libel post was Conditional Fee Agreements (CFAs, or “no win no fee”).  They were brought in to provide access to justice for all, but as I explained they provide huge incentives for law firms to get you to exploit bad laws and impose a huge cost upon defendants.  It’s hard to see how one could spuriously defend a legal action, it’s not possible to make success fees conditional on the size of the award (there isn’t one!), and access to legal representation is even more important when defending (economic disparity is key to this), so CFA reform in my regime would probably leave it as-is for the defence.

That said, while it is possible to defend against SLAPPs using a CFA (one example here), it seems less usual because of discovery costs, drawn out proceedings and unlikelihood of winning.  Possible remedies for this include capping discovery costs and mandating that plaintiffs provide evidence early in the process (IANAL so I really can’t speak to how to do either effectively); another is to provide something similar to California’s Code Of Civil Procedure 425.16, which provides a special motion defendants may file in the case that the suit infringes upon their First Amendment rights.  Unfortunately we don’t have the US Bill Of Rights (you can keep the Second Amendment though, the whole firearms thing is a disaster) so we’d have to come up with our own wording – on the plus side, this would allow us to broaden it’s scope to explicitly include things like creativity and enterprise.

Penalties for SLAPPers

I’ve already mentioned CFAs, and I’d add remedial awards (loss of earnings and whatnot) to those, but neither is meant to be punitive.  Allowing for such, particularly in conjunction with an early counter-motion statute as described above, would be a serious disincentive to filing such suits.  Given that the essential problem is the economic disparity when (for instance) a corporation sues an individual, I would have this as unlimited and at the judge’s discretion, although obviously in a common law system so heavily enamoured of precedent as ours we’d need to keep a watchful eye on this.

The Streisand Effect

The good news on this front is that the nature of the internet, social media and ubiquitous reporting is undermining the SLAPP as an effective weapon.  The Streisand Effect describes the phenomenon of attempted censorship serving to draw much more attention to the information in question – a great recent example is the British Chiropractic Association’s action against Simon Singh which led to the criticism being widely reported in the press and the web, over 500 chiropractors having formal complaints of false advertising levied against them, and the spurring on of the campaign for libel reform.  We have a long way to go with this – for instance, the Trafigura scandal and accompanying superinjunction might have been broken by widespread coverage on Twitter, but the follow-up on the issue has been extremely disappointing, especially in the mainstream press.  Still, I feel there is cause for optimism, and as I’ve said before the worst culprit – the UK’s libel laws – look like they are in for a change.

A quick primer for those not familiar with defamation or libel: defamation is a statement or statements that makes a claim about a “named or identifiable individual (or individuals) in a manner which causes them loss in their trade or profession, or causes a reasonable person to think worse of him, her or them” (from Wikipedia).  Libel is written and published defamation.  In the age of the internet, I would like to think that libel laws will become unnecessary – the net should be a place where everyone has the right to reply and have their voice heard, so any falsehoods should hopefully be exposed and the perpetrators suffer a blow to their reputation and creditability; however, this utopia does not yet exist, and it would be naive in the extreme not to admit that certain information sources (newspapers, large news sites, TV channels) have disproportionate influence and that adequate redress is possible simply by refuting them.  We therefor still require some form of legal framework for libel claims, but the one we have at the moment in the UK is fundamentally flawed in a number of ways, and so I propose the following reforms.

Radically overhaul the way CFAs work

In 1998, the Courts and Legal Services Act was expanded to include the use of Conditional Fee Agreements (“no win no fee”) to defamation cases, thus lowering the barrier to launching a libel suit.  In theory, this allows access to justice for those who couldn’t otherwise afford it; in practice it also makes libel suits as a means of silencing criticism much more attractive.  It’s also a huge incentive for legal firms to engage in libel actions, because if successful (including a paid settlement as well as a final judgement) they can claim all billable hours plus a “success fee” of up to 100% directly from the defendants – and this is not conditional at all upon the size of the judgement, so as you can imagine having even the smallest libel judgment made against you can be financially crippling – Lord Justice Jackson is on record as saying these fees are sometimes over 1000% of the award or settlement, and study from the University of Oxford found the costs of libel proceedings in the UK to be 140 times higher than the European average (direct link: PDF warning).

While Jack Straw is looking to cap CFA success fees at 10% of the billable hours, the way to correct this imbalance is to link the success fee directly to the size of the award or settlement.  This would encourage lawyers to focus on libel cases where actual damage has occurred (for more on this, see below).  Fortunately, this is the one area that looks like it will be addressed in the near future.

CFA reform goes further than simply libel, by the way, partially addressing the abundance of personal accident claimants and huge associated costs that are the real cause of one of the Daily Fail’s favourite boogeymen, the elf-n-safety-brigade.  I’ll be coming back to this in a future post.

Introduce redress for the defense

Even if you successfully defend a libel case, avoiding paying damages and the CFA costs, you’re going to be substantially out of pocket, as unlike the claimant you have to pay your fees up front, and you’re unlikely to get it all back.  For instance, see Ben Goldacre’s clash with pill salesman Matthias Rath, which cost The Guardian £170,000 and 19 months of struggle (and, not inconsequentially for a newspaper, being prevented from commenting on the story).  Defendants should be able to claim back all costs; additionally, changes I propose below should result in more cases being thrown out sooner which will also keep costs down.

Of course, it’s better all round if we have fewer cases to start with, which brings me on to the real meat of UK libel law today.

Remove the Duke of Brunswick rule

A ruling dating back to 1848 established that if a libelous piece was reprinted that counted as a fresh publication and, as such, reset the timer on the statute of limitations on defamation action.  Due to the way web servers work – sending a new copy of a page each time someone requests it – these are covered by that same rule, meaning that for any piece of writing on the internet liability for potential libel action expires not after 6 years, as was the intention of the law, but in fact persists as long as the document is accessible – in reality, that means forever.  It also means that you can take legal action anywhere in the world the site is available, as viewing it there is an act of “publishing” – since the UK has notoriously strict libel laws, as we’re discovering, this fact has led to a glut of  libel tourism here.  Further, as anyone familiar with the internet will tell you, it’s more or less impossible to make something disappear once it’s out there.

This rule is archaic and ridiculous.  It needs purged entirely.

Remove liability from service providers

Service providers who exercise no editorial control (or remove content retroactively) could be held to be ‘publishers’ and therefor liable.  Quite aside from it being ridiculously unfair to hold someone liable for someone else’s content they did not approve or condone, this creates two serious problems: firstly, there is a chilling effect on the cultural and technological innovation that has thrived on the open platform of the internet, and services such as hosted blogging simply could not exist were the providers required to manually scan (presumably with legal counsel) each entry to avoid lawsuits; and secondly, it allows the circumvention of due process if one wishes simply to have content removed – threaten the provider with legal action, and you might well be able to get content you disapprove of removed without the knowledge or intervention of the alleged libeler.

Section 230 of the US Communications Decency Act

No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.

I don’t see that it should to be limited to “interactive computer service” (“information content providers”, maybe?) but I’d introduce something very close to this.  Like CFA reform, it would have be introduced as a separate item to my libel reform bill, as it’s much more widely applicable than just defamation, as I’m sure I’ll cover in future WICTP posts.

Put the burden of proof back onto the claimant

At present, claimants have to prove that the allegedly libelous writings refer to them and that they are defamatory, which is to say they suffered a loss of trade/profession or that a “reasonable person” would think less of them (including being subject to ridicule, contempt or hatred).  This means that the burden of proof is on the defendant to prove that their comment was either true or justified.  I’ll come to standards of justification shortly; for now let’s focus on the burden of proof.  At present, we are essentially valuing people’s reputation more than their freedom of speech, and as a result have low requirements to begin pursuing a libel action.

The solution to this is simple: upon bringing the suit, the claimant should be required to demonstrate that the claim is false – or in the event they cannot do so provide adequate reasons for this – and unjustified.  If they fail to do this it should be tossed out, and I’d also like to see punitive damages for suits brought with the intent of stifling valid criticism (such as the Trafigura scandal) – again, this is something I mean to explore in greater depth in subsequent WICTP posts, although the use of the web and social network to neuter that ruling is heartening for the future.

Require claimants to demonstrate actual damages

While actual malice must be proven to claim punitive damages for all claimants and compensatory damages in the case of public figures, private individuals need only prove negligence to receive compensatory damages (and, as we’ve seen, as soon as they’re awarded the legal fees kick in).  Moreover, claimants are not required to demonstrate that they have suffered to the extent of the damages they are claiming, meaning a blogger with 10 readers could be subject to the same awards as a newspaper with millions of readers.

To remedy this, I’d increase the standard for private individuals to malicious defamation rather than negligence (where malice includes repeated failure to correct oneself if one has been shown to be in error), and assert that the claimant must demonstrate the actual damages for which they are claiming.

Expand public benefit justifications

Currently, there is no broadly applicable public benefit defence against libel.  There is an established public benefit justification in the form of the Reynolds defense, which allows for cases where a journalist had a duty to publish an allegation even if it turns out to be wrong, essentially turning any trial in which it is deployed into an investigation of the journalist’s conduct as well as the content itself.  While I’ve no reason to think it wouldn’t be available to all, the criteria which is adheres to are very much based on the traditional model of journalism (such as seeking comment from the plaintiff and verifying sources).  These things aren’t necessarily applicable to, for instance, correct and proper scientific inquiry, which is clearly a public benefit, as is demonstrated by the continuing actions brought against scientists and scientific journalists by proponents of unproven disciplines, such as Simon Singh.  It also doesn’t account for situations where the steps of the Reynolds defence can’t reasonably be followed, such as reporting on a totalitarian dictatorship.

The remedy to this is similar to the issue of requiring malice to be demonstrated.  I would create a public interest defense for any statement made that was believed to be both true and in the public interest, and place the burden of proof against these upon the claimant as described above; I would also add an explicit defense for science and scientific journalism, requiring that the principles of scientific inquiry were followed (to be adjudged by independent expert witnesses, preferably practicing scientists) and that results are not deliberately misrepresented or used to make unrelated defamatory claims.  This would necessarily be a low bar – the correct way to respond to scientific criticism is with science, not with law.

Expand fair comment justifications

The overriding problem this section faces is that of context – both in terms of what the defendant intended to say, and whether in context we should consider it to be opinion rather than fact.  For the former, one need only look as far as Simon Singh’s case again to see the problem – reading his entire article (as you can here), he clearly didn’t intend to say the British Chiropractic Association were deliberately deceptive, and yet that is the meaning he was instructed by the judge to defend.  Fortunately his case is currently in appeal, but the principle of context needs codified in law to prevent this from ever happening.

For the latter, fair comment is a slippery devil.  What exactly is factual? Everyone makes statement that, out of context, sounds entirely factual; must we be required to prefix everything with “I think that”, like a bad high school essay, in order to prove we’re dealing with opinions?  Or perhaps we could enclose everything in quotations to prove it (I’m looking at you, BBC News).  Then there’s the argument that by stating exactly what is fair comment, you declare everything you don’t explicitly mention not to be, which is likely to cause issues in the future.  My suggestion would be to firstly put the burden of proof upon the claimant to show the offending statement(s) are intended as factual, and secondly to apply judgment to the situation – the “reasonable person” standard seems appropriate, seeing as it’s what we’re supposed to be using in defamation law.

So, where to?

In the real world where I’m not in charge (sigh), where do we go from here?  Well, first I’d suggest reading LibelReform.org’s report covering much of the ground I have here in more depth; then, if you’re a UK voter, please consider this issue when voting.  The Conservative Party, unfortunately but predictably, don’t appear to want to take this bludgeon away from the wealthy and powerful (incidentally the letter in that link is a form response, I received one too and I have a different MP); Labour we’ve already discussed; and the Liberal Democrats appear to have genuinely got it.  Regarding the smaller parties, the Greens only mention blasphemous libel as far as I can tell, UKIP are too busy being overly nationalistic to worry about such things, and the Pirate Party are discussing it, and it definitely fits with their (our, actually, as I’m a member) principles, but there’s no official position yet.

I’ll keep a full list of posts here, along with a summary of the changes I’d make.

Libel Reform

• Overhaul Conditional Fee Agreements and make them contingent on the award
• Introduce stronger measures of redress for a successful defence
• Abolish the multiple publication rule
• Remove liability from service providers
• Move burden of proof onto the claimant rather than the plaintiff
• Require demonstration of actual damages
• Expand ‘public benefit’ defence
• Expand ‘fair comment’ defence

I should state straight-up that I’m assuming, off the back of purely anecdotal data, that there is a higher percentage of atheists, skeptics and freethinkers in the computer science fraternity than the general population (correcting for affluence, education levels and whatnot).  If anyone has hard data on atheism or skepticism demographics by profession (outside of academic science and philosophy, which have excellent stats) I’d love to see it, but for now let’s plow ahead with my assumption and ask what the potential factors are here.  Please also bear in mind these are generalisations and tendencies – there are most definitely exceptions to these.

We spend our lives debugging

Debugging is the act of stepping through a flow – be it code, an argument or a real-life situation – to look for discrepancies; things that aren’t as expected, whether it’s the state or the logic.  The parallels to philosophy are obvious – an argument is essentially a piece of simple (by CS standards) logic, which is either valid or invalid, that draws a conclusion from a premise or premises, such as: (premise) All men are mortal; (premise) Socrates is a man; (conclusion) therefor Socrates is mortal.  We’re built to spot logical inconsistencies, because in our world, they’re bugs that must be squashed.

As a corollary, I strongly suspect we’re much more susceptible to certain fallacies – incorrect premises, straw men (see below)  – than strict logical errors, such as affirming the consequent or denying the antecedent.  I also agree with Zed Shaw that programmers often overvalue logic and undervalue evidence.

We are generally meritocratic

Software engineers, in general, don’t value authority for authority’s sake.  We’re logical, and we value problem solving above niceties – that’s what we love to do, and we recognise it in others, as well as appreciating those who make our lives easier (The Truth About Managing Geeks is a good read that touches on some of this).  The upshot is we’re less likely to take the word of some authority figure at face value (aka the fallacy of argument from authority) – we’ll evaluate what we’re told, decide whether it’s worthwhile and ignore it if it’s not (and, more than likely, subvert it if it’s forced upon us – there’s more than a little anarchist in many a techie I know, including myself!).

The flipside of this is arrogance.  We can be dismissive of those who don’t share our technical knowledge, and assume that because we’re knowledgeable in our own profession those skills automatically transfer to other domains – or worse, that those domains are inherently inferior to our own and not worthy of consideration.  This is by no means limited to computer scientists, as the age-old philosophy vs science debate will attest, but it’s worth keeping in mind.

We are opinionated and confrontational

Any programmer who’s interacted with the community at all is going to be familiar with the Holy Wars.  Topics as seemingly innocuous as whether to use Java over C++, or how to format code most readably, create massive arguments, histrionics and flame wars.  The problem I have with these is that too often both sides have good points, are completely unable to see the points of the other side, and descend into completely over-the-top ad hoc rationalisations and name calling.  Sad, but true – and the reason I suspect we’re more susceptible to straw men than other fallacies.

I think in general we’re accustomed to speaking out in no uncertain terms when we have a strong opinion on something, which garners attention that quite possibly contributes to the impression of skepticism I have (or maybe it’s just confirmation bias on my part, who knows).  If so, on the whole it’s a good thing, although lacking receptiveness to others’ points is something to avoid.

We are often anti-convention

The picture of the geeky outcast living in a basement shunning all contact save fellow geeks is overly harsh, but there lies a kernel of truth at the heart of this stereotype – I certainly wasn’t in with the popular crowd at school, and I know I’m not alone in that.  Western culture can be very anti-intellectual at times, and compscis come in for some of the worst of it.  The result is often a latent mistrust of what is popular or mainstream.  When married to the ability and desire to validate claims made to us on their own merits, this insulates us from argumentum ad populum; the all-too-often seen result, though, is simply supporting a position because others don’t (for instance, the Dissent From Darwin petition – which I refuse to link to – features a strangely high number of computer scientists, although that may in part be down to the Salem Hypothesis).

Any thoughts?  This is something of a mind dump, so is likely laden with too much projection and confirmation bias – hence why I’d love to see some hard stats on the subject, or at least some other opinions.

Let me first clarify what I mean.  In terms of an actual chart battle, clearly this wasn’t representative; in fact, given the various ways we came up with to play the system most effectively – in particular pushing the easiest/cheapest places to get the track, as well as a means to download it multiple times for free and have it count – I’m reasonably sure that more people bought the X-Factor track than bought RATM, and the difference was made up by multiple purchases (I bought more than 1 copy).  I don’t doubt more people than were involved in the effort dislike the recent predictability and the populist blandness Cowell et al inflict upon us, but this wasn’t about a numerical advantage.

This is significant as it means the defining factor in the victory was the Facebook campaign.  I know a bunch of my FB friends (and people I follow on Twitter) were pushing the campaign, and I plugged it once too.  It’s been described as an organic campaign, which is kind of true.  The couple who started the campaign (Jon and Tracy Morter) certainly weren’t pushing this to make money, although they had started a similar campaign the previous year to get this ditty to #1.  Using that experience they grew the Facebook group that was the nexus of the campaign to huge numbers with a little help from folks like Peter Serafinowicz (275k followers) and others plugging it.  It spread like wildfire leading up to the first day of the week that counted, and everyone was coordinated to buy it then, gaining plenty of attention in the mainstream media.  There was also a final push on the Saturday, as it transpired that RATM was behind, and it seems to have dramatically increased the sales then.

So what do we learn from this?  Well, first off, grassroots social campaigns work.  It certainly helps to have a little experience in running them – and they do need run, they don’t just happen by themselves.  Having a cause and a target date you’re working towards allows you to build towards a crescendo, which differs from the release and slow burn small-starting viral campaigns I’ve seen around newer content, but you need to really strike a cord with many many people to make that work.  It’s still not clear to what extent mainstream exposure should be a end goal of such campaigns – clearly enough sales were obtained on day 1 to get attention, but it’s hard to say if victory would have been possible without it.  There are still players that are disproportionately influential, and that’s always (to my knowledge) as a result of a pre-existing fanbase – which makes sense, as Facebook/Twitter are communication platforms first and foremost – and this influence is valuable, particularly if you’ve built an actual relationship with your fans (if and how quickly one would lose such influence were it abused is still an open question).  Finally, being able to accurately track the state of a social campaign very much influences its effectiveness, so measurement will continue to be hugely important.

With regards to X-Factor, I suspect the whole reality show model will continue to be the major popular music vehicle going forward and as music sales become less relevant to the industry.  People are clearly willing to pay for the level of involvement that the phone-in vote gets them, it’s the only profitable, unpiratable Connect With Fans + Reason To Buy model the existing labels have come up with, and it fits the same build-to-crescendo style of marketing as the RATM campaign.  I’d expect to see even more of it going forward, but hopefully now other artists won’t run in fear of providing an alternative while it’s going on.

Replace the ‘final’ keyword with the inverse, ‘mutable’

Mutable state is a necessary evil.  Immutable objects are simple to create and use, are automatically thread-safe and make testing much less painful, amongst other things.  Moreover, the fact that classes are overridable by default encourages people to use inheritance rather than lovely testable composition for code reuse.  Unfortunately though, nothing in Java is immutable by default, so you have to make a concious choice to do the simpler, better thing.  It’d be much better to have to declare a member or class mutable so that developers had to explicitly do it wrong.

Allow labelled parameters

public void drawSquare(int ~width x, int ~height y) {}
drawSquare(~width: 2, ~height: 3);
drawSquare(~height: 3; ~width: 2);

That way I don’t have to go and look up the API to see which way round they have to go.  Sure, leave the defaults the way they are – this clearly would have to be optional (maybe throwing a compile-time error if they’re wrong rather than swapping them?) – but it could help nip nasty logic bugs in the bud.

Allow default parameter values

public void foo(String bar = "douglas") {}

beats

public void foo(String bar) {}
public void foo() {
  foo("douglas");
}

in my book.  Syntactic sugar, but brevity and readability win out.

Implement arrays properly

byte[] a = {1, 2, 3};
byte[] b = {1, 2, 3};
a.equals(b); // will return false

Why?  Oh, I can use Arrays.equal()?  Fantastic, static calls, my favourite.  Oh, and there’s the whole nested array thing too, just for extra fun.

Allow switch statements on Strings

switch(myString) {
  case "foo":
    // some stuff
    break;
  case "bar":
    // other stuff
    break;
  default:
    // even more stuff
}

is more readable than

if ("foo".equals(myString)) {
  // some stuff
} else if ("bar".equals(myString)) {
  // other stuff
} else {
  // even more stuff
}

Syntactic sugar again – and also a possible source of bugs given casing and the Turkey problem – but I’d still like it.

Fix generics

Enough is written about this elsewhere.  Suffice to say, losing generic information at compile-time is a whole world of suck (it’s why you can’t have primitives as generics without boxing, for instance).

Swing should be banned.

Ok, I lied about ignoring libraries, but it cannot be said often enough that Swing is pure distilled evil.

The miscommunication happens at least in part because of the conflation of moral facts – what we reason to be good or bad – with moral intuition – what we feel to be good or bad.  The latter is trivially relativistic – people produce different moral judgements on different situations, as it witnessed by the fact that people who commit what we would consider great evils often feel morally justified in doing so.  They are also influenced in this by social and cultural circumstance – slavery, to pick an oft-used example, was considered morally acceptable by many more people in 18th century America than almost anywhere in the world today.  Our moral intuition is an evolutionary artefact of the benefits of group living, an inclination to act to benefit the group as a whole, and so obviously varies from individual to individual, and with the dynamics of said group.

However, if we are to exist in complex, organised societies, we require some means of determining moral facts that apply to all within that society.  This is where we often find a disconnect between the atheist and the theist, where the theist claims that if there are no moral facts and each person is free to make their own morality.  This is technically true of moral relativism*; it is, however, highly misleading, as it puts one in mind of moral anarchy where there is no consensus and no ‘good’ or ‘evil’ beyond what one personally approves or disapproves of.  It implies that as we cannot make an absolute moral judgement, we cannot make any moral judgement on more than an individual level.  An obvious example of a relativist moral system that lets us make such judgements is social convention – a moral fact is defined as a moral intuition that a majority of the group agrees upon, and if the intuition of the majority changes, so too does the moral fact.  This is something I feel more popular moral relativist atheists need to emphasise – allowing the moral realist vs moral relativist discussion to be framed in absolutist terms will do the atheist cause no favours.

The second disconnect is the assumption of many theists that objective morality may only be derived from a supernatural source (God).  William Lane Craig words this as a logical argument for the existence of God: 1) if objective moral facts exist, God exists; 2) objective moral facts exist (his supporting evidence appears to be “and I think we all know it”, but we’ve addressed that already); 3) therefore God exists.  It is 1) that is most problematic for me.  As it happens I am a moral realist – I believe absolute, objective moral facts exist, and that they are not dependent on any supernaturalism.  The moral theory I subscribe to is Alonzo Fyfe’s desire utilitarianism – where morality is defined in terms of desires (which we know to exist, making this an entirely naturalist theory).  A moral desire is one that tends to fulfil other desires, and an immoral desire is one which tends to thwart other desires.  Diverging into why I subscribe to this theory takes us outside of the scope of this post, but if you’re interested a more in-depth introduction can be found here; my point is merely that it, as well as objectivism, humanism, some forms of philosophical libertarianism and others, there exist theories of objective moral facts that are not dependent upon a supernatural power.

So those are the two points I feel atheists should emphasise when discussing morality: that moral relativism does not mean moral anarchy, and that moral realism is not dependent upon a god or gods.

*Amusingly, the idea that moral behaviour is defined as acting entirely in your own self-interest - egoism – is generally held to be a realist position.  I rather suspect that what most theists who rail against relativism (particularly in the USA) actually object to is pluralism – the idea that multiple conflicting moral facts are all valid in their own contexts.  We wouldn’t want to prevent the pious judging of other cultures, would we