How recording self declared gender of offenders instead of sex will distort statistics

Murray Blackburn Mackenzie (MBM) argue for, and shared a petition for, the accurate recording of rapists’ sex so as to avoid distorting crime statistics. In a response to a Freedom of Information request, Police Scotland stated they would record gender via self id, and thus if the male self identified as a female the crime would be recorded as a female offence, hence MBM’s petition here.

When I retweeted the MBM petition on Twitter, Steve Dennis (who has since blocked me) claimed that it would not make a difference because the definition of rape requires the rapist to have a penis, thus you can infer if a female is recorded as having committed a rape, then that crime was committed by a biological male.

It is true that if you see crime stats showing females committed X rapes in a given year you can infer those women were males self id’ing as females because of the way rape is defined. However this does not mean there isn’t scope for distorting the statistics:

• Just because the inference can be made does not mean it will be made. This is not a case of splitting hairs:-
  • the point really only holds where rape statistics are considered in isolation from other offences.
    • Suppose the police databases contain a table of recording each crime committed and the gender of those who committed it. Someone viewing the table may decide this is the right table to query to get a total number of offences broken down by sex, write the query to simply count how many offences had female recorded against them vs male, and believe the job is done. They would thus have included the rapes in the total without being aware of it. Even someone who is aware of it, may forget (perhaps under pressure to meet a deadline) to make the adjustment, or might make a mistake in implementing the adjustment.
    • any statistics that these numbers will be fed into will be distorted unless the people who produce the derived or aggregated statistics are aware of the need to make adjustments.
    • there is thus plenty of scope for distortions to arise as the original context of the figures is lost and the need to make the inference is forgotten about, or even as people try to make the adjustment.
  • if there are constant reports about female offending that include rapes in the figures, even if breakdowns are provided so that someone could make the inference, people will tend to only look at the headlines (leaving the breakdowns as detail) and the female propensity to commit crime will be exaggerated in people’s perceptions.
• The inference can *only* be made in the case of rape.
  • If you use self declared gender for all offences, then the only offence for which you can then determine the sex from these records becomes rape, and only because of the way it is defined. For all other offences, if someone declares themselves to be female or male, this will be recorded and there’ll be no way of deriving actual sex from the crime records alone. You’d need to merge the records with e.g. birth certificates and medical records to make the inference.

Sex based policies are not inherently trans exclusionary.

In a recent discussion on Twitter, it was put to me that sex-based¹ policies are inherently trans exclusionary.

Whilst I can see how inappropriate, inconsistent or insensitive use of sex-based policies may cause harm to trans people (and possibly others), the claim that a sex based policy is inherently trans exclusionary seems obviously incorrect to me because the decision to include/exclude is made based on sex and will both include (trans) people of the specified sex and exclude (trans) people not of the specified sex and do so without regard to trans status.

If sex is the relevant determining factor in the decision, it is quite legitimate to use it for the decision and the question then is how to do so sensitively and in a manner consistent with protecting the interests of the person concerned.

I consider this issue in more detail below.

UK bill proposes making it an offence to cause ‘serious annoyance’ to someone without ‘reasonable excuse’.

Update: The bill’s 2nd reading is imminent (scheduled for 15th/16th March).

Boris Johnson’s government seems to have decided to crack down on those who cause serious annoyance to others. The latest Police, Crime, Sentencing and Courts Bill has been recently introduced in the House of Commons.

According to section 59 of the bill, a person commits an offence if the person does an act (or omits to do an act they’re required to do by law) and the person’s act or omission causes ‘serious harm’ to another person.

For the purposes of the Act, a person is deemed to have suffered serious harm, if they suffered:

• (a) death, personal injury or disease,
• (b) loss of or damage to property,
• (c) serious distress, serious annoyance, serious inconvenience or serious loss of amenity, or
• (d) were put at risk of suffering anything mentioned in (a) to (c)

It is a defence to have a reasonable excuse for committing whatever action it was that caused the ‘serious harm’.

Deeming points (a) and (b) as serious harm is entirely reasonable and I’m not fussed about that, but the inclusion of ‘serious annoyance’ is another matter entirely. Who will decide whether a given level of annoyance is ‘serious’? The person claiming to be annoyed? The arresting officer? The CPS? The court? How can this be determined?

What evidence can a complainant give that they were seriously annoyed by someone’s actions beyond angrily stating “I was seriously f*cking annoyed!”?

Will cartoonists have to think up reasonable excuses for drawing unflattering caricatures of our politicians, celebrities, or members of the royal family?

Note that your action need not actually cause someone serious annoyance, it need only be deemed to risk causing someone serious annoyance. Under this bill, if you have any reason for thinking anyone might be offended by something you intend to say or do then you have to consider whether your reasons for saying or doing it will stand up in court.

Understanding the “trans” debates: what it means to be “cisgender” and why some object to the term

In my earlier article on what it means to be transgender, I explained some of the terminology that Stonewall and many other trans rights activists use with respect to trans people. Naturally they also have terminology for describing those who are not transgender. Non-trans people are often referred to as “cisgender” (being the opposite of “transgender”) or “cis” for short. So if you’re a non-trans woman, you’d be referred to as a cis woman. Similarly to their use of “trans” they regard “cis” as an adjective/modifier.

Stonewall defines “cisgender” and “transgender” as:

Cisgender or Cis

Someone whose gender identity is the same as the sex they were assigned at birth. Non-trans is also used by some people.

…

Trans

An umbrella term to describe people whose gender is not the same as, or does not sit comfortably with, the sex they were assigned at birth.

From the Stonewall glossary, version at time of writing archived here.

Note that this terminology assumes that we all have a gender identity and that our sex is something assigned to us at birth. This has led to many people objecting to the use of the term “cis” e.g. on the following grounds:

• they do not have a gender identity. They may further object that there is little or no scientific basis for the idea that we all have a gender identity i.e. an innate sense of our own gender.
• sex is observed at birth, not assigned and it relates to our reproductive biology and is thus a fact about our bodies, not a subjective feeling of maleness or femaleness.
• the term carries the suggestion that a non-trans person is comfortable with the gender associated with their sex when they may actually be gender non-conforming but have no problem with being of the male or female sex, or being seen as male or female.
• many non-trans women object that agreeing to use the term “cis” to describes themselves means agreeing to terminology that demotes them to a subclass of what was previously their own class when the terms woman/women referred to their sex. I.e. they object to the redefinition of woman/women to include biological males. Note that many trans rights activists will view disagreement that trans women are women as a form of transphobia in and of itself. Indeed Stonewall’s definition of transphobia reads: “The fear or dislike of someone based on the fact they are trans, including denying their gender identity or refusing to accept it. Transphobia may be targeted at people who are, or who are perceived to be, trans.”

What we see here then is a clash between people who define terms such as male/female and man/woman on the basis of biological sex/reproductive anatomy, and those who define them based on gender identity. It is also a clash between those preferring a definition based on objectively observable traits and those preferring a definition based on internal feelings/subjective sense of identity.

The 2020 US election is not quite over yet

As of the time I write this, counting is still going on, however that is not the reason for my post as I don’t expect completion of the count to alter the Biden win that’s been indicated so far. With Trump refusing to concede, the question boils down to whether the electoral college will vote Biden in. It seems clear to me that Trump’s tactics are targeted at the possibility of winning at this stage of the process.

With the results in so far, it is clear that Biden has won the popular vote in states with enough seats in the electoral college to win the Presidency, assuming the electors vote according to who won the popular vote in their respective states. The margins are too high for either the remaining uncounted votes or any recounts to alter the situation. This is the basis on which all the major news channels have called the election for Biden, and why Biden himself is claiming victory and making preparations to enter the White House.

In most US elections, the opposing candidate would have conceded by now and thus the assumed victory would be confirmed. However Trump has not conceded. He and his supporters claim that the election was rigged, whilst stories are appearing in the media about alleged voting irregularities in the swing states where the margins of victory are narrowest. The Trump campaign has started lawsuits to challenge aspects of the way the election was run and/or to advance claims of voter fraud.

Many people may be assuming that the Trump campaign’s actions are just the actions of a defeated President trying to cling to office, and that whilst perhaps a few irregularities might be found, there won’t be anything found on the scale required to overturn the result. This not like Bush vs Gore in the year 2000 where Bush won the narrowest of victories in Florida, with the final count giving a majority of just 538.

In the 2020 election however we’re talking about wining margins of multiple 10s of thousands of votes and it looks like at least 3 states results would need to be overturned. It seems unlikely to me that fraud on a scale sufficient to flip the results in the required states would be found, and on that basis one might conclude that Biden will become the next President regardless of the Trump campaign’s tactics here.

However, it seems to me that what Trump is aiming for isn’t necessarily to get sufficient Biden votes discounted to overturn the results in the swing states (though of course that would be ideal for him), but to cast sufficient doubt on the results in those states that it gives the electors in the electoral college a reason to depart from voting for the candidates who won in their state.

The electors are not legally bound to vote for the candidate who won in their state, doing so is simply a convention in the US system that has been honoured for most elections. If court rulings find that there has been fraud targeted at the swing states, and that fraud goes beyond a trivial amount, doubt can be cast on the legitimacy of the results in those states without necessarily invalidating sufficient votes to flip the results. If many thousands of votes in each state are proven to have been fraudulent in court cases, the reasoning will be that the proven fraud may just be the tip of the iceberg that’s been found by investigations so far, ergo we can’t be sure who won in the respective states, even if the proven fraud is not yet sufficient. Extra weight for this argument might also be provided if the legal actions have also prevented certification of the results. Would the electors for those states then decide that actually they should vote for Trump after all? I’m not sure we can rule it out in this scenario.

Time will tell whether this gamble will work for Trump and secure him the Presidency. Even if it doesn’t, the exercise is still casting doubt on the legitimacy of the vote in people’s minds and this may cause problems for the Biden presidency as a result.

Why I am sceptical about the UK’s “moonshot” proposals for mass covid-19 testing

The UK government has proposed regular mass testing for COVID-19 (the so-called “moonshot” proposals) so that people who test negative for the disease can return to something resembling pre-COVID life. To do so will require a different kind of test for whether someone is infected with the SARS-CoV2 virus (which causes COVID-19) to the tests we have so far.

There are several things to note about the current tests — the figures below are based on this BMJ article and running scenarios through the calculator there:

• the lower end(*) of the estimates of sensitivity (aka recall) is at 70%. This means that for every 100 infections the tests will return a positive result for 70 of them. There would thus be 30 false negative results.
• the lower end of the estimates for specificity is ~95%, meaning that for every 100 non-infections they will return a negative result for 95 of them, but will incorrectly return a positive result for 5 of them. There would thus be 30 false positives.
• they do not test for whether someone is infectious. They only test for whether they are infected, i.e. for presence of the SARS-CoV2 virus.
• the probability that someone who tests positive actually has the disease depends on the prevalence of the infection in the population they’re drawn from.
  • If you select randomly from a population where only 1% of the population is infected, this would translate to a ~16% probability that someone who tests positive is actually infected.
  • However if you were to draw from a population where say 80% have the disease, a positive result would entail a ~98% probability of having the disease.

(*) the article suggests using the lower end of these estimates due to there being high variability in the estimates and the likelihood that existing studies are overestimating sensitivity.

The numbers above demonstrate that these tests would not be suitable for mass testing of millions of people every day. For example, if we assumed that 10% of the population were infected at any one time (well above the levels reached for COVID-19 so far) then for every 1 million people tested:

• 100,000 people would be infected, of which 70,000 would test positive and 30,000 would falsely test negative.
• 900,000 people would not be infected, of which 855,000 would test negative and 45,000 would falsely test positive.

We’d thus be telling tens of thousands of people per day they’re infected when they’re not. We’d also be telling tens of thousands of people they’re not infected when they are. Trust in the system would quickly vanish.

Note that at the moment, the government advises people to get a test only if:

• they experience certain symptoms, such as a new continuous cough, fever or loss/change in sense of taste and smell, or
• live with someone who has the above symptoms, or
• are otherwise asked to do so by the authorities or
• are taking part in a pilot project.

They are thus selecting from groups of people who are suspected of or at risk of having the disease already or testing people in the context of research. In these contexts, the test is already accurate enough for the purpose of confirming a suspected case of COVID-19 or the test itself is being assessed for some purpose in a pilot project.

The question arises of whether it is possible to devise a test that could be used in the way the government requires. They posit people running tests at home that then produce a result and if it’s negative it means those people can go about life as if they could know for sure that are not infectious. I.e. we need a low false negative rate, much lower than the 30% rate of the current tests (which do not indicate infectiousness in the first place). However focusing on getting a zero false negative rate could mislead in that you can achieve a zero false negative rate simply by deciding that the whole population is infectious!

We thus need to have a test that can genuinely detect infectiousness for the SARS-CoV2 virus and where a negative result indicates a high probability (e.g,. 95%+) of not being infectious – thus we optimise the test for the purpose of correctly detecting when people are not infectious rather than for detecting the presence of the virus (as the current tests do). It also needs to be a rapid test that can be done repeatedly and done at home.

I don’t know how feasible it is to produce such a test, but I note that the BMJ are highly critical of the “moonshot” idea in this article. They point out that we’d need to do repeated mass screening of the population for this approach to work, and that there is a risk of forcing large numbers to self isolate unnecessarily due to the risk of false positives arising during testing. They further note that the studies suggesting use of rapid tests oriented towards screening have all been based on computer modelling rather than actual studies of screening programs.

This all makes sense to me – in my own field of machine learning, when faced with a classification task there is often a trade off to be made between low false positive rates and low false negative rates and if you optimise for the low false negative rates then you may get high false positive rates.

Understanding the ‘trans’ debates: What Stonewall and many others mean by the term “transgender”

Readers may be aware of the often very heated debates, occurring in both social media and more traditional channels, about issues such as gender self identification, the use of single sex facilities and allowing trans women to compete in women’s sports, all of which revolve around the rights of transgender people and rights that have, until recently, been associated with a person’s biological sex.

This article focuses on the terminology used by many of those who advocate for trans people.  My aim here is to explain this terminology in the hope it may help people new to the issues to understand the debates should they get involved.

In the remainder of the article, I set out how Stonewall, one of the most influential LGBT groups in the UK, defines the term transgender (often shortened to “trans”) and draw out some implications.

I chose Stonewall because they are well known and because they are clearly influencing both the debates and the policies surrounding transgender rights and the terminology is broadly in line with what I’ve seen in the debates.

Stonewall’s definition of trans(gender)

In this section I present Stonewall’s definition of “trans”, “transgender man” and “transgender woman” plus some further terms related to what it means to be transgender from their glossary of terms (definitions below taken from version archived here):

Trans

An umbrella term to describe people whose gender is not the same as, or does not sit comfortably with, the sex they were assigned at birth.
Trans people may describe themselves using one or more of a wide variety of terms, including (but not limited to) transgender, transsexual, gender-queer (GQ), gender-fluid, non-binary, gender-variant, crossdresser, genderless, agender, nongender, third gender, bi-gender, trans man, trans woman,trans masculine, trans feminine and neutrois.

 

Transgender man

A term used to describe someone who is assigned female at birth but identifies and lives as a man. This may be shortened to trans man, or FTM, an abbreviation for female-to-male.

 

Transgender woman

A term used to describe someone who is assigned male at birth but identifies and lives as a woman. This may be shortened to trans woman, or MTF, an abbreviation for male-to-female.

 

Non-binary

An umbrella term for people whose gender identity doesn’t sit comfortably with ‘man’ or ‘woman’. Non-binary identities are varied and can include people who identify with some aspects of binary identities, while others reject them entirely.

 

Gender dysphoria

Used to describe when a person experiences discomfort or distress because there is a mismatch between their sex assigned at birth and their gender identity.
This is also the clinical diagnosis for someone who doesn’t feel comfortable with the sex they were assigned at birth.

 

Gender identity

A person’s innate sense of their own gender, whether male, female or something else (see non-binary below), which may or may not correspond to the sex assigned at birth.

 

Homosexual

This might be considered a more medical term used to describe someone who has a romantic and/or sexual orientation towards someone of the same gender. The term ‘gay’ is now more generally used.

 

Key points about the Stonewall definitions

Here I note several points about the definitions above that are important in understanding where they and many other trans rights advocates are coming from:

• Sex is described as being assigned at birth. Many trans advocates use this terminology and thus talk as if one’s biological sex is something that is assigned to you rather than being a fact about your body.
• Stonewall distinguishes here between sex and gender –  clearly with “trans” people Stonewall are referring to someone’s sex not matching or at least being somewhat at odds with their gender.
• The examples in the definition of “trans” make it clear that “trans” is broad enough to describe anyone from those who simply cross dress to those who decide to have surgery (transsexuals) to look like a member of the opposite sex.
  • It’s beyond the scope of this article to define all of the terms used in the list of examples given under the definition of “trans”. For now I simply note that the term non-binary covers everything that falls outside the male and female genders.
• You could define a trans woman as a person whose gender identity is female but whose sex recorded at birth is male, a definition which is also compatible with the definition given in the glossary whilst in my experience being more in line with how these terms are used in practice.
• The definitions of a transgender man and a transgender woman make no reference to surgery, hormone therapy or gender dysphoria. From the glossary’s definitions of these terms and that of being trans itself, it is thus clear that none of these are a necessary part of being trans. 

Implications

Below are some implications of defining things according to the above terminology. These implications are the main sources of contention in the debates:

• Intact biological males can be trans women without any surgery or other modification of their bodies.
  • An example is Alex Drummond (from Stonewall’s trans advisory group) who in this video rejects having surgery.
• Similarly, biological males can thus also be lesbians under this terminology, e.g. in this interview published on BuzzFeed, Alex Drummond identifies as a lesbian.
  • Stonewall has defined homosexuality / being gay in terms of gender, not sex:- a trans woman and a non-trans woman are both considered to be of the female gender.
• It follows that if use of e.g. women’s facilities is to be determined by gender / gender identity rather than sex, then these facilities will no longer be single sex.
  • It is often stated in these debates that “trans women are women”, and that “trans” serves as an adjective here, cf “short women are women”. (“Trans men are men” and “Non binary identities are valid” may also be stated, but this happens less often in my experience. )
• One cannot tell someone’s gender / gender identity simply by looking at them. This point is stated explicitly in this guide to being a trans and non binary ally from the Trevor project (a charity supported by Harry Potter actor Daniel Radcliffe). To know whether someone is male, female or non-binary you have to ask them.
• The demand that Stonewall and other trans advocates make for gender self identification thus follows logically on from the terminology and the implications set out above.

Do the UK covid-19 guidance & regulations permit Dominic Cummings trip to Durham?

People are calling for Dominic Cummings to be sacked after reports that he and his wife and 4 year old child traveled by car from London to Durham, whilst he had symptoms of Covid-19. Both Cummings himself and a Downing Street spokesman have since defended the trip. According to the BBC report, the statement from Downing Street reads:

“Owing to his wife being infected with suspected coronavirus and the high likelihood that he would himself become unwell, it was essential for Dominic Cummings to ensure his young child could be properly cared for.

“His sister and nieces had volunteered to help so he went to a house near to, but separate from, his extended family in case their help was needed. His sister shopped for the family and left everything outside.

“At no stage was he or his family spoken to by the police about this matter, as is being reported.

“His actions were in line with coronavirus guidelines.”

Cummings has also claimed his actions were reasonable and legal.  So, do the guidelines allow for this?

Let’s consider a case that some are making that they do. This web page of government guidance tells people to stay at home and self isolate if symptomatic, but it appears to have a get out clause:

If you are living with children

Keep following this advice to the best of your ability, however, we are aware that not all these measures will be possible.

Additionally Dr Jenny Harries, Deputy CMO for England, in this clip from today’s press briefing on the coronavirus, when asked about this situation, does say that if the parents of a child are unable to care for that child and can’t get support locally then this would count as an exceptional circumstance. Defenders of Cummings are suggesting thus that the “exceptional circumstance” applies and justifies Cummings decision to drive his wife and child to Durham where extended family could then support them.

An additional piece of evidence being cited is the lockdown regulations themselves which supply a non-exhaustive list of “reasonable excuses” to leave home that includes this reason in section 6 (2) (d): “to provide care or assistance, including relevant personal care within the meaning of paragraph 7(3B) of Schedule 4 to the Safeguarding of Vulnerable Groups Act 2006, to a vulnerable person, or to provide emergency assistance;”

So lets assume that the account given in the Downing Street statement is accurate. It would thus appear that Cummings anticipated the possibility of he and his wife becoming sufficiently ill to be unable to look after the child, and took his family to a residence in Durham, near his extended family who could provide support during the self isolation, including looking after the child should that become necessary. In doing so, he took the risk of spreading the disease if for any reason they had to make contact with others during the journey (e.g. if they stopped at services for a bathroom break for their 4 year old child).

The defence of this action is that the “exceptional circumstance” Harries mentioned, and the get out clause in the guidance and the lockdown regulations themselves all permit this.

For this to be the case, it seems to me that both of the following conditions must apply for this to fit with the guidance:

1. Cummings and his wife were unable to find friends, relatives or other support in London to deal with this possibility.
2. Cummings extended family were unable to come to London, even to take the child back to Durham to be cared for whilst the parents were ill, let alone stay in London to provide support.

Remember Cummings and his wife were both symptomatic and decided to drive 250 odd miles, risking spreading the disease if they had to stop along the way, so for this to be reasonable in the circumstances (to trigger the “exceptional circumstances”) you pretty much have to assume they had no other realistic choice than the action they took. That seems unlikely to me.

For it to fit with the lockdown regulations themselves means we must either interpret the circumstances as being an unlisted “reasonable excuse”, or interpret provision of care to the child to cover temporarily moving to a different residence so that the person who could provide the care if both parents fell ill would be able to do so. Maybe one of these could be argued in a court of law by a skilled lawyer but it is not obvious to me that this interpretation would stand unless the 2 conditions above applied.

In the meantime, I expect there will be numerous cases of families where both parents were worried about falling ill but stayed put because of the government’s guidance and instruction to do so, who will be looking at this situation with incredulity and anger.

Letter to MEPs regarding Articles 11 and 13 of proposed copyright directive

Dear MEPs,

I am writing to urge you to vote against the proposed EU copyright directive when the plenary vote is held on the 5th July, due to the inclusion of the proposals in Articles 11 and 13, which I believe represent a major threat to freedom of speech on the internet, will undermine development of open source software and will put European internet companies, especially small companies, at a serious competitive disadvantage. 

Regarding Article 13, this makes platforms that allow users to upload content liable for any copyright infringements those users might make, unless they can demonstrate they made best efforts to prevent the upload appearing in the first place. 

 

This will mean that simply in order to minimise the chances of being held liable for an infringement, they will have to check every single upload against copyrighted works to judge whether they infringe those works. This will entail automated filters being applied to all uploads of content whether it be music, photos, video, text, podcasts or even code (e.g. on sites such as Github), and the platforms will err on the side of caution to block uploads if there is any doubt.  

This is a seriously misguided approach to copyright enforcement for the following reasons:

• Such automated filtering requires general mass surveillance of user’s activities on the internet and thus constitutes a general monitoring obligation, which is on contradiction to rulings of the European Court of Justice. 
• Algorithms cannot tell whether copyright has been infringed. 
  • They can at best indicate that there is some content in common shared between an uploaded file and any files in the database being checked. However this does not tell you whether the copying is permitted under fair use laws or even if whether it’s the uploaded file or the file in the database that’s violating copyright.
  • No filter can possibly review every form of content covered in the mandate including text, audio, video, images and software.
    • Article 13’s mandate is technically infeasible.
    • Attempts to automate filtering of videos and music, such as Google’s Content ID, are already making ridiculous errors and applying this to every platform and every form of copyright is simply insane.
  • Users will thus find their freedom of expression curtailed because it will be difficult to quote from or to parody copyrighted works, even though it is perfectly legal to do so, due to the filters over-blocking.
• The requirement to put such filtering in place will add considerably to the cost of setting up any internet based businesses that involve enabling users to share content of any sort and thus put European start-ups and other European countries at a major disadvantage,
• The requirement to filter all forms of copyrighted material this way impacts on code sharing sites and thus will undermine the development of open source software that has powered much of the modern internet, including most major AI and machine learning platforms and packages, as explained in this post from Github (probably the most widely used code sharing site currently). Note that clauses restricting the requirement to not for profit code sharing sites do not address these concerns properly.   

Regarding Article 11, my understanding is that this will require people to get a licence if they wish to quote snippets from journalistic content (e.g. newspaper and magazine articles) in online works. This again is a throughly misguided proposal:

• Quoting from articles is normally protected by fair use laws and is standard practice in academic discourse, blogging, and for that matter in journalism itself.
• The proposal does not even allow a publisher to waive the requirement for a licence thus interfering with the rights of those who wish to publish under the Creative Commons licence. 
• The proposal violates the Berne convention.
• The aim of the proposal is apparently to address the issue that some publishers believe that news aggregators, by automatically providing titles and/or snippets when linking to articles are somehow depriving  the publishers of those articles of revenue. This fundamentally misunderstands what is happening here:
  • By providing links with snippets to news articles, the news aggregators drive traffic to the sites concerned, that is they provide publicity for the site/articles concerned and consumers will visit who would not otherwise have done so. Attempts to prevent people from providing such links will simply lead to those sites being ignored.
    • As this article explains, the operation of similar proposals in both Germany and Spain failed and led to many publishers rejecting them.  
  • There are numerous ways a site can prevent aggregators from picking up articles from (sections of) their website. For example, to stop Google (and many others)  indexing pages you can simply set up a robots.txt file. See also: http://www.robotstxt.org/

I believe both of these proposals, if enacted, will do great harm to the openness of the internet in Europe, will seriously undermine freedom of speech on the European internet, and will put European companies, especially the smaller companies, at a major disadvantage as they try to deal with the burdens these would impose on them.

I urge you please to vote against the copyright proposal accordingly, in the upcoming plenary vote on July 5th.

Yours Sincerely,

James Hammerton

On leaving the EU with no deal: Potential disruption to air travel

In her Lancaster House speech, Theresa May said:

And while I am confident that this scenario need never arise – while I am sure a positive agreement can be reached – I am equally clear that no deal for Britain is better than a bad deal for Britain.
(emphasis added)

Given this position, it’s worth knowing what the consequences of “no deal” would mean. In the worst case scenario (e.g. talks breaking down) this means looking at the consequences of Britain reaching the Article 50 deadline without any sort of agreement whatsoever having been struck with the EU.

In this scenario, on the day before Brexit, Britain will still be a full member of the EU enjoying all the privileges and subject to all the obligations that entails. On the next day, Britain will be outside the EU completely, treated in the EU’s parlance as a “third country” and having to deal with the EU on the default terms they apply to any non EU/non single market country. When references are made to a “cliff edge” in the media comment on Brexit, it is this sudden transition from full EU member to being outside the EU (and the single market) that is being referred to.

Here, I consider the consequences of leaving without a deal for the airline industry, where recently the Guardian reported that UK airlines may have to open bases of operation in other EU states and/or sell off shares to EU nationals post Brexit or lose major routes.

I conclude that if there is literally no deal, the likely consequence will be considerable disruption to UK/EU airline industry. This is in neither the UK or EU’s interests, though and a deal that ensures the UK remains part of the regulatory system for aviation in the EU should be feasible.