Hacker News Reader

Comment by isk517 | Hacker News Reader

My question is out of those 2,498 people stopped how many of them actually had links to terrorisms? I will be extremely generous and say if the number is less than 2000 than sounds like you just have a system in place to harass random people.

View on HN

For imbalanced data, even accurate system will give tons of false positives. Unless you are willing to tolerate false negatives, having large false positive rate is almost inevitable. Rare disease testing is a popular example, but this one is a good instance to I think.

Let's say there are 10 million people linked to terrorism in the world and the system is correct 95% of times. Then out of 20000 people, 20 are positives, and 19 will be stopped; out of 19980 negatives, 999 will be stopped.

Compared to disease testing you may argue that FP is more harmful, but that's a different question. The point is that even very non-random system will be dominated by FP.

Edit: in particular, if they stop truly random people and 500 of them are linked to terrorism that implies that 20% of all people passing there are terrorists.

> Unless you are willing to tolerate false negatives

We tolerate false negative all the time: Terrorism is not eradicated at all in UK. Barring successful results, the police should stop harassing people.

I don't think we do tolerate false negatives, for this class of thing, all the time.

Terrorists passed checkpoints today and we tolerated it. Terrorists passed checkpoints yesterday and we tolerated it. Terrorists passed checkpoints the day before that...

The complete elimination of false negatives for rare occurrences is prohibitively expensive to the liberty of non-terrorists.

Being unable to eradicate something is not the same as to tolerate it.

Probably important to note that this leads to tests not being performed without good reason because they cause more harm than they do good due to the false positive rate.

Great question. The journalist named in the article (Matt Broomfield) is asking authorities to delist the PKK (which clearly IS a terrorist organization targeting civilians) as a terrorist organization.

https://twitter.com/MattBroomfield1/status/15092064663512760...

That's a stretch to say linked to terrorism, unless the Belgian supreme court is as well.

Source on PKK targeting non-paramilitary/police/military civilians in the last 20 years?

e: I can find this from 1987 (a time when the Turkish state was also actively involved in massacring civilians) https://en.wikipedia.org/wiki/P%C4%B1narc%C4%B1k_massacre

To name a few:

* December 2016 Istanbul car bombing / suicide bombing near a football stadium, killing 44 people

* October 2016 Istanbul bombing - injuring 10 civilians

* June 2016 Istanbul bombing / killing 6 civilians

* March 2016 Ankara car bombing killing 37 people

* June 2006 Manavgat (in the province Antalya, a popular region for tourists) detonation of a bomb in a tourist site, killing four people

Sources:

https://en.wikipedia.org/wiki/October_2016_Istanbul_bombing

https://en.wikipedia.org/wiki/June_2016_Istanbul_bombing

https://www.dni.gov/nctc/ftos/pkk_fto.html

> * December 2016 Istanbul car bombing / suicide bombing near a football stadium, killing 44 people

TAK != PKK, TAK claimed responsibility. 39 of those killed were police officers. Turkish military also kills civilians while targeting opposing militia groups.

> * June 2014 Istanbul bombing / killing 6 civilians

Link? Did this happen?

> March 2016 Ankara car bombing

TAK not PKK, again.

> * June 2006 Manavgat

Yet again, TAK.

> https://en.wikipedia.org/wiki/October_2016_Istanbul_bombing

Not intentional targeting civilians, accidental premature explosion of bomb. Only one of the attacks you've mentioned that was by the actual PKK and actually killed civilians. (e: actually no civilians killed)

To his defence, the TAK is often claimed to be a militant wing of the PKK, that handles affairs that might damage the reputation of the main organization. Like how the Provisional IRA formed fake splinter groups to pin sectarian counterattacks on.

Western government have had high-profile negative media coverage in the past because they weren’t keeping track of returnees from Syria, my guess is they are being more diligent in response to that.

Regardless of the morality it is unusual and you aren’t really supposed to go to Rojava.

Considering they asked him questions like "do you consider your reporting objective" and didn't return his phone and laptop, it's gone a bit beyond counter-terrorism. I would be less heavy-handed towards a journalist, due to the importance of a free media

I suspect the overwhelming majority had "links" to terrorism, for a broad enough definition of "link."

Intelligence services are good at stopping and intimidating journalists who are reporting on their activities: activities like nurturing Islamic terrorist groups in order to destabilize states perceived as an obstacle. For examples see the 9-11 attackers, the Ariana Grande/ManchesterArena bomber.

They have no lack of information: they just don't do anything with it because their fundamental job at the moment is not protecting us, it's destroying others. MI5 had all they needed to stop the Manchester Arena bomber etc.

Journalists need to be free to pursue their activities in order for us to have a functioning democracy. The detention of Kit Klarenberg, this detention, the detention of the French publisher (for Gilet Jaune protestor links IIUIC) are definite indications that these intelligence agency perverts feel empowered to interfere with one of the important parts of a democratic system.

Everyone has a link to terrorism and Kevin Bacon, if you dig deep enough.

Oh, there's nothing random about them.

Reply on news.ycombinator.com