There is a trade-off, though—if you suddenly remove a large certificate vendor, you risk training users to click through the warnings, and end up potentially worse-off than you were before. The more recent trend seems to have been to require the use of Certificate Transparency by issuers that have been caught misbehaving.

At the moment we are just about the only one that doesn't—no liquid limits on domestic flights, and no restrictions on non-passengers coming airside.

That isn't the case though---the decision whether or not to publish is made before the copyright agreement is even raised, and the agreements always include provisions for US government works (which aren't copyrighted) and Crown copyright (which are not generally signed over).

In terms of figuring it out, how simple do you think this agreement is on paper? Six lines in the middle of an A10 sheet with room for big signatures?

In this case, not really that far off: http://eur-lex.europa.eu/legal...

Article 107

(ex Article 87 TEC)

1. Save as otherwise provided in the Treaties, any aid granted by a Member State or through State resources in any form whatsoever which distorts or threatens to distort competition by favouring certain undertakings or the production of certain goods shall, in so far as it affects trade between Member States, be incompatible with the internal market.

[Various exceptions].

NASA has been involved in lots of classified stuff, as I understand, over the years—their kind of rockets aren't all that different from missiles after all.

As far as export control goes, there is a lot of stuff on the export control lists; radation hardened computers, and most anything that can take stuff into space and/or back, among many many other things.

The actual multiplications are nowhere near as fast. A multiplication of an RSA-sized number takes thousands of cycles (see here), and modular arithmetic of that size is even slower. 44kHz corresponds to a sample per 45k 2GHz cycles, and Montgomery multiplication as in the link above takes up to two adds per bit if you do it quickly and insecurely, with each taking on the order of 100 cycles. An exponentiation of a 1024-bit message will need therefore around 100k (average-case) cycles i.e. 2.5 audio samples. This will go increase at least quadratically with key size, meaning that with 2048-bits you're looking at ten samples on average.

In any case, they are a reputable bunch, you'll notice Shamir (the S in RSA) in the author list.

Note: I am not associated with the Tor project, just an interested observer. I happen to be implementing a similar protocol for something else.

Because it needs to be resistant to compromised nodes. The reason for this that hidden service connection details (though not the server IP obviously, since all of this happens through Tor channels) are stored in directory servers which are randomly assigned each day. The choices of directory server are derived from a pseudo-random string [1]

descriptor-id = SHA1(permanent-id | SHA1(time-period | descriptor-cookie | replica))

by taking taking hashes of the directory identity details and sorting, and then picking those that come after descriptor-id in the list.

The problem is that a malicious would-be directory can modify its own configuration so that its hash changes in order to gain responsibility for an arbitrary hidden service at some point in the future, since the descriptor-id values are predictable. This doesn't give them complete control, but they could perform DoS and traffic counting.

What was proposed last year, then, was to add a random element to the data being hashed so that it could not be predictable [2]. In order to prevent there being a single point of failure (both from a reliability and security point of view), it was proposed to use a distributed random number generator. The way that this works is that while the master directory servers agree on the list of relays, they also generate a random value and use a bit-commitment protocol [3] to commit to it before the final value is generated in order that the last server to vote can't just keep generating random values until it finds one that gives it control of a given service.

The way that this happens, then, is that during the first half of the day the directories will include committed values with their votes on the network status. During this time everyone should get a copy of the committed value, which is generated by hashing a random string [2]. Then, during the second half of the day, they reveal their chosen random values. The others can then hash the received value and compare it with what they were given before in order to make sure that they have not changed their random value in response to the other random values.

At the end of all this the revealed values get hashed together in a particular order and the resulting value is published and put into the descriptor-id by server operators and clients. You can't use one of those idQuantique etc. cards and call it a day because there's nothing to stop a compromised server from emitting random values that are favourable to an attacker, whereas this approach will still be unpredictable so long as at least one of the master directory servers is honest and takes part.

[1] Tor Rendezvous Specification
[2] Tor Proposal 250: Random Number Generation During Tor Voting
[3] Commitment scheme

Sure, but it's not necessarily competing with R. There are a _lot_ of Matlabbers who don't use C interface yet still run uncomfortably lengthy simulations, and for them Julia is hugely advantagous. Static typing would be something of a turnoff when you're coming from something like Matlab where you don't think about types at all in most programs, especially faffing about in the REPL.

Australia is no liability, no ifs, no buts, unless you were negligent---negligent is defined as being similar to such things as telling people your PIN, or writing it down undisguised and storing it with your card.

Looking at the judgement, it seems that the issue was that TripAdvisor claimed repeatedly in their advertising that the reviews were true, genuine, and trustworthy, but that the investigator was able to post blatantly false reviews. From footnote 146,

A titolo meramente esemplificativo si riporta il testo di alcune di tali recensioni:

i) “Ci è piaciuto tantissimo!!! Ma non sono sicuro se era questo ristorante o el kebab che è lì vicino. I filtri di TA non funzionano qui si può scrivere qualsiasi cosa”, recensione rilasciata per il ristorante “Combal.zero” di Rivoli e pubblicata in data 6 settembre 2014;

ii) “I’ve never been here!!! This websites has NO filters so I can say anything about this Restaurant and everyone is going to believe it. Buonanotte”, recensione rilasciata per il ristorante “Osteria francescana” di Modena e pubblicata in data 6 settembre 2014;

iii) “È senza dubbio il miglior ristorante cinese di Milano. Ottima l’anatra, gran buffet, camerieri gentili. Fantastici filtri sulle recensioni come potete osservare! Cinque palle verdi”, recensione rilasciata per il ristorante “Pomodoro & basilico” di San Mauro Torinese e pubblicata in data 4 settembre 2014.

[Probably terrible] translation:

i) We liked it _so_ much! But I'm not sure whether it was this restaurant or the kebab shop nearby. TA's filter doesn't work...here one can write whatever they want

iii) It is without doubt the best Chinese restaurant in Milan. Excellent duck, big buffet, polite staff. These are fantastic filters of the reviews, as you can see! (note: the restaurant is named "Tomato & Basil" and so clearly not Chinese)

We have a distinction between summary and indictable offences here, the cutoff being two years maximum in gaol.

The issue was that because people weren't getting headhunted the companies didn't need to compete as heavily on wages. If I recall correctly, when one of the deals fell through, Google compensated with a substantial payrise, which was used to justify the scale of the loss.

I recall reading that when Emirates started putting in mobile phone facilities, the staff on the plane reported that they didn't hear anyone using them, whereas the logs showed lots of calls being made. The engine noise was such that the phone calls are essentially unnoticable.

The hijacking that you were thinking of (Ethiopian Airlines Flight 961) had survivors. According to that article, a significant number of the dead were those who didn't wait until leaving the plane before inflating their lifejackets, and as a result were trapped inside the plane when it flooded. 125 out of 175 died, with only 60 to 80 still in their seats.

It's also not done in Australia either. The process here is that everyone puts their university/course preferences into a website. Then, at the end of the year after exams are marked (identical across the state), students are ranked and the system goes down the list and assigns people to university places. No wishy-washy essays or personality assessments, you get a number based on your marks and that determines what you get.