One of the nice things about bad movies is that, once you know the name, you can avoid that movie again, very easily, by not seeing a movie of the same name.
However, when bad movies are stripped of their crap component and some video Dr Frankenstein sews them back together with thin thread of little plot, it becomes harder to avoid them.
Welcome to Doomsday. Escape From New York meets Mad Max: Beyond Thunderdome meets Waterworld meets a bad rip-off of Braveheart meets 28 Days Later meets the Raymond Revuebar meets ... oh, you get the idea. Remember that boring and stupid crowd scene from the Matrix II? There's an even dumber one here. Rubber-bound gimp from Pulp Fiction? Check! Strange tip of the hat to Edge of Darkness? Confusing but it's here. Gratuitous product placement of Bentley motorcars that is probably a reference to Cherry 3000? I keep wondering how much more they can jam in - it's not surprising that there's no plot. This is a fever dream of unrelated set pieces with the same actors handling similar roles. It's like "Thank God You're Here" with an R Rating and people exploding in clouds of fluid.
Our poor good guys can't get a break. Everywhere they are, hundreds of people have been mysteriously waiting, apparently for the last 25 years, to jump them. With large buses being run on the fuel that somehow survived 25 years of isolation. Given that the stations in Adelaide start to run dry after 3 DAYS without resupply, I'd love to know how big the storage tanks are in Scottish service stations. Let's not even talk about the complete recreation of feudalism, complete with armour, depicted as being superior to gunpowder. Hmm...
This film is awful. Mindless gore, without real humour. Farcical science fiction that doesn't stand up to any questioning. This is probably the worst film that I've seen recently - and I've seen every film Milla Jovovic has ever made...
Let's not make any hasty comparisons here, but there is no point in seeing this movie if you've been keeping up with science fiction films in the past couple of decades. They're all consumed here, digested partially and then regurgitated in one long disconnected set of chunks.
This is a terrible film. They manage to make an exploding gimp boring. Don't bother.
Saturday, October 04, 2008
Monday, September 29, 2008
Three Days of Sunday
I missed a dinner engagement tonight with some very nice people. I realised at 8:30 that I should have been at dinner at 6:30pm. The reason that I forgot was that I didn't know that today was Monday.
I realise now that, on Saturday when I had people coming over to play games, I had thought it was Sunday. A friend and I were convinced that he would miss the final episode of Doctor Who because it was on that night. No, it wasn't. Then, the next day, it was Sunday again. I was lying here with a leg in the air, with a brief diversion for a very nice lunch. But it was Sunday again.
Today, I didn't go out, I watched TV, I read books. I lay on my back with my leg in the air again. I'd been sending e-mail backwards and forwards all weekend about work and the flow of work mail today didn't trigger my Monday sense.
Fortunately, my friends are patient and nice - and they're having me to dinner tomorrow night instead.
I have put a note on the door, with the time and the day and the name of the people I'm seeing.
Fingers crossed.
I realise now that, on Saturday when I had people coming over to play games, I had thought it was Sunday. A friend and I were convinced that he would miss the final episode of Doctor Who because it was on that night. No, it wasn't. Then, the next day, it was Sunday again. I was lying here with a leg in the air, with a brief diversion for a very nice lunch. But it was Sunday again.
Today, I didn't go out, I watched TV, I read books. I lay on my back with my leg in the air again. I'd been sending e-mail backwards and forwards all weekend about work and the flow of work mail today didn't trigger my Monday sense.
Fortunately, my friends are patient and nice - and they're having me to dinner tomorrow night instead.
I have put a note on the door, with the time and the day and the name of the people I'm seeing.
Fingers crossed.
Friday, September 26, 2008
Global Finance
Hey, economists!
Is it possible to have a steady-state economic situation that avoid recessions, given that indefinite expansion is impossible, or are you all just making stuff up while high on nose-candy?
I still can't get my head around the current meltdown in the US. I can understand the creation of something out of something else, except that if the words 'until house prices start dropping' occur to me after examining things for a few seconds, shouldn't this have occurred to other people?
Of course, I'm being naive. The secret of finance is to hand on the hot potato before it burns you. There are a group of people who have made vast amounts of money from what has happened and a lot of this money will continue to be worth something, even if they plugged it into the stock market. The really big companies knew that, despite the dogma of contemporary capitalism, the US government would be forced into a semi-socialist buyout and funding round to support the economy. All the benefits of the invisible hand, except that it only has one finger extended.
What do I know? I'm the wrong kind of scientist. Any economists, fill me in and inform me.
Is it possible to have a steady-state economic situation that avoid recessions, given that indefinite expansion is impossible, or are you all just making stuff up while high on nose-candy?
I still can't get my head around the current meltdown in the US. I can understand the creation of something out of something else, except that if the words 'until house prices start dropping' occur to me after examining things for a few seconds, shouldn't this have occurred to other people?
Of course, I'm being naive. The secret of finance is to hand on the hot potato before it burns you. There are a group of people who have made vast amounts of money from what has happened and a lot of this money will continue to be worth something, even if they plugged it into the stock market. The really big companies knew that, despite the dogma of contemporary capitalism, the US government would be forced into a semi-socialist buyout and funding round to support the economy. All the benefits of the invisible hand, except that it only has one finger extended.
What do I know? I'm the wrong kind of scientist. Any economists, fill me in and inform me.
Thursday, September 25, 2008
Italy and Knee Surgery
I never post while I'm actually away somewhere - I don't like people to know that I'm not at home, given that I'm using my own name on this blog.
I've recently returned from Turin, Italy, where I attended a conference and workshop group to present and chair a session. Very successful trip and I hope to write up a more complete report shortly. In short, Turin is beautiful. The food and wine are great and the city is a postcard, surrounded by small villages that are what good postcards hope to be when they die - Barolo, Barbarseco, Alba, beautiful!
I am currently sitting on the couch at home, recovering from arthroscopic knee surgery. Before the surgery, neither the doctor nor the surgeon were very optimistic about me running again for any real distances. However, the surgeon has now cleaned up my knee cap and trimmed back some tissue that was getting pinched in the knee and, assuming that the healing goes well, he seemed much more positive about running in the future. That's great news. I can always take up swimming, I guess, if things go pear-shaped. My ligaments, cartilage, minisces and bones are actually in good shape - no arthritis or tears. Hooray!
I hope to post more often in the near future. Watch this space.
I've recently returned from Turin, Italy, where I attended a conference and workshop group to present and chair a session. Very successful trip and I hope to write up a more complete report shortly. In short, Turin is beautiful. The food and wine are great and the city is a postcard, surrounded by small villages that are what good postcards hope to be when they die - Barolo, Barbarseco, Alba, beautiful!
I am currently sitting on the couch at home, recovering from arthroscopic knee surgery. Before the surgery, neither the doctor nor the surgeon were very optimistic about me running again for any real distances. However, the surgeon has now cleaned up my knee cap and trimmed back some tissue that was getting pinched in the knee and, assuming that the healing goes well, he seemed much more positive about running in the future. That's great news. I can always take up swimming, I guess, if things go pear-shaped. My ligaments, cartilage, minisces and bones are actually in good shape - no arthritis or tears. Hooray!
I hope to post more often in the near future. Watch this space.
Sunday, June 22, 2008
Thoughts on Information and Torture
Hi,
I've covered some of this before on my old LJ, as well as in comments to other people, but I thought I'd put it together in one neat package.
Assume that you have a very simple language, composed of three symbols {plants, trees, are}. All combinations are valid sentences. Fairly obviously, the set of all possible sentences from this language will be small and will consist of all the strings of 1, 2 and 3-symbols long. If order is important, as it is in this language, we distinguish between 'plants are trees' and 'trees are plants' as being distinct sentences. Thus, the total number of permutations available are P(3,1) + P(3,2) + P(3,3), 3 + 6 + 6, which is equal to 15.
If we assume that the observer of the formed sentences is able to use any or all standard punctuation, or rules for handling sentence fragments, to add an external meaning to the language sentence then we can consider the following to be legal sentences.
'plants.'
'trees.'
'plants are.'
'are plants?'
'trees are.'
'are trees?'
'plants are trees.'
'trees are plants.'
'are trees plants?'
'are plants trees?'
(Feel free to argue for any of the others!)
In other words, 10 of the 15 possible sentences can, through random selection, form a sentence that could either be a question or an answer with meaning in the context of some arbitrary observer.
What does this mean? It means that, by chance alone, I have a 66% probability of giving the fictional observer information that they may consider to be sensible (or semantically grounded, even if illogical).
One of the issues with torturing someone for information is that, if they don't actually know anything, it is possible for someone to accidentally produce a set of symbols that are significant for the observer. The above, simple, example shows the possibility. Of course, the actual universe of discourse is enormous and there are vastly more elements at play in English than the three that are observed here. I hope, however, to illustrate how this supposedly vast array of symbols can be limited through interrogation techniques to the stage that random assembly is a sufficiently great risk to undermine any other information received.
If someone had kidnapped you, put you in a locked room and subjected you to a range of disorientation techniques, you would be confused, scared and disorientated (quelle surprise!). If someone then asked you a series of open-ended questions such as:
"What were you planning to do?"
"Who are your accomplices?"
"Where did you hide it?"
then the chances of you answering such a question, even if you wanted to, with a false random positive is incredibly small. You have no framework to make the story believable, as even an assumption of terrorism is questionable at this stage. Should the interrogator start torturing you, and you have no real knowledge, your survival instinct and pain avoidance mechanisms will start to dominate. Ultimately, there is a risk that you will say anything to try and save your own life. Without any context, your babblings will probably make no sense. Consider this alternative questioning technique:
"When was the bombing planned? When the shopping centre was full?"
"We know you're working with Mr X (name denoting ethnic background), who else are you working with?"
"Where did you put the bomb? Where did you get the drums from?"
You now have, in effect, a little language composed of key symbols of value to your interrogator.
Assuming that the truth has failed you, and that the interrogators have now moved to the next stage and are torturing the crap out of you, it is quite possible that you will confess to anything to save your remaining hand, or your eye, or to just not drown. Under duress, many people would be tempted to assemble something in order to prevent further harm to themselves. Synthesising something like "Mr X brought in this guy, he said it was his cousin, his name was 'Y (ethnically similar)'. I don't know where the drums came from. The bombing was scheduled for 4pm."
For those who doubt this, please read up on the witchcraft trials and how a number of people were tortured until they confessed to witchcraft, at which point they were executed. Knowing that certain death awaited them, the victims eventually confessed to escape further pain and to them move onto death. The vocabulary in this case was much easier, any combination of words leading to a pact with the devil or a simple signature could suffice.
Under the effect of torture, I have little doubt that many people will hand over all of the information that they know. It is, however, impossible for an interrogator to know when this point has occurred and if faulty information, or simplistic profiling, has been used to grab a suspect, the interrogator may not believe the suspect when they claim ignorance. Under the effect of torture, people will admit to just about anything to get the torture to stop - including farcical claims about deals with the devil, when the only reward is death.
Thus, if we regard torture as a function that maps a person's knowledge to their utterances (and label it as T), for a person with knowledge (k), we assume that T(k) produces U(k), the set of truthful utterances possible from k. However, U(k) may still be generated, without knowledge k, as the symbol set of language is finite and the likely introduction of key symbols through the interrogation process. Thus, a person without knowledge !k, can also produce U(k), albeit with a lower likelihood.
Ideally, if someone is going to torture someone else, they must know some of the true sentences, which I'll label as check sentences C(k), contained in U(k), in order to validate the output from T. The question is, how many of these do we need in order to determine that it is unlikely that U(k) has been generated randomly? If poor questioning technique or leading questions are employed, then U(k) would have to be vast. Given that U(k) is vast, how many sentences in C(k) will we need to ensure that we know when a valid sentence in U(k) is uttered?
And here is the crux of the matter: given that we don't know U(k), we cannot guarantee that C(k) is a reliable and indicative subset of U(k). If C(k) is unreliable, then none of the sentences in U(k) can be validated unless they can be separately validated through a reliable external mechanism. For those who don't know, heresay evidence doesn't really count here and, unfortunately, heresay evidence is used extensively to provide C(k). More importantly, if C(k) is used as a source for questions applied during the interrogation and there exists a set P(k), which are legitimate permutations of C(k) that exist within U(k), then an unknowledgeable participant can provide a credible superset of C(k) that appears to enhance the interrogator's knowledge of U(k).
And then this randomly generated knowledge becomes a fact, for the purposes of the interrogators, and may be used against the same person again, or on other people. It wouldn't surprise if real criminals have actually been let go because they had no knowledge of the largely fabricated scenarios that are generated by other people who have been pulled into the net.
I've covered some of this before on my old LJ, as well as in comments to other people, but I thought I'd put it together in one neat package.
Assume that you have a very simple language, composed of three symbols {plants, trees, are}. All combinations are valid sentences. Fairly obviously, the set of all possible sentences from this language will be small and will consist of all the strings of 1, 2 and 3-symbols long. If order is important, as it is in this language, we distinguish between 'plants are trees' and 'trees are plants' as being distinct sentences. Thus, the total number of permutations available are P(3,1) + P(3,2) + P(3,3), 3 + 6 + 6, which is equal to 15.
If we assume that the observer of the formed sentences is able to use any or all standard punctuation, or rules for handling sentence fragments, to add an external meaning to the language sentence then we can consider the following to be legal sentences.
'plants.'
'trees.'
'plants are.'
'are plants?'
'trees are.'
'are trees?'
'plants are trees.'
'trees are plants.'
'are trees plants?'
'are plants trees?'
(Feel free to argue for any of the others!)
In other words, 10 of the 15 possible sentences can, through random selection, form a sentence that could either be a question or an answer with meaning in the context of some arbitrary observer.
What does this mean? It means that, by chance alone, I have a 66% probability of giving the fictional observer information that they may consider to be sensible (or semantically grounded, even if illogical).
One of the issues with torturing someone for information is that, if they don't actually know anything, it is possible for someone to accidentally produce a set of symbols that are significant for the observer. The above, simple, example shows the possibility. Of course, the actual universe of discourse is enormous and there are vastly more elements at play in English than the three that are observed here. I hope, however, to illustrate how this supposedly vast array of symbols can be limited through interrogation techniques to the stage that random assembly is a sufficiently great risk to undermine any other information received.
If someone had kidnapped you, put you in a locked room and subjected you to a range of disorientation techniques, you would be confused, scared and disorientated (quelle surprise!). If someone then asked you a series of open-ended questions such as:
"What were you planning to do?"
"Who are your accomplices?"
"Where did you hide it?"
then the chances of you answering such a question, even if you wanted to, with a false random positive is incredibly small. You have no framework to make the story believable, as even an assumption of terrorism is questionable at this stage. Should the interrogator start torturing you, and you have no real knowledge, your survival instinct and pain avoidance mechanisms will start to dominate. Ultimately, there is a risk that you will say anything to try and save your own life. Without any context, your babblings will probably make no sense. Consider this alternative questioning technique:
"When was the bombing planned? When the shopping centre was full?"
"We know you're working with Mr X (name denoting ethnic background), who else are you working with?"
"Where did you put the bomb? Where did you get the drums from?"
You now have, in effect, a little language composed of key symbols of value to your interrogator.
Assuming that the truth has failed you, and that the interrogators have now moved to the next stage and are torturing the crap out of you, it is quite possible that you will confess to anything to save your remaining hand, or your eye, or to just not drown. Under duress, many people would be tempted to assemble something in order to prevent further harm to themselves. Synthesising something like "Mr X brought in this guy, he said it was his cousin, his name was 'Y (ethnically similar)'. I don't know where the drums came from. The bombing was scheduled for 4pm."
For those who doubt this, please read up on the witchcraft trials and how a number of people were tortured until they confessed to witchcraft, at which point they were executed. Knowing that certain death awaited them, the victims eventually confessed to escape further pain and to them move onto death. The vocabulary in this case was much easier, any combination of words leading to a pact with the devil or a simple signature could suffice.
Under the effect of torture, I have little doubt that many people will hand over all of the information that they know. It is, however, impossible for an interrogator to know when this point has occurred and if faulty information, or simplistic profiling, has been used to grab a suspect, the interrogator may not believe the suspect when they claim ignorance. Under the effect of torture, people will admit to just about anything to get the torture to stop - including farcical claims about deals with the devil, when the only reward is death.
Thus, if we regard torture as a function that maps a person's knowledge to their utterances (and label it as T), for a person with knowledge (k), we assume that T(k) produces U(k), the set of truthful utterances possible from k. However, U(k) may still be generated, without knowledge k, as the symbol set of language is finite and the likely introduction of key symbols through the interrogation process. Thus, a person without knowledge !k, can also produce U(k), albeit with a lower likelihood.
Ideally, if someone is going to torture someone else, they must know some of the true sentences, which I'll label as check sentences C(k), contained in U(k), in order to validate the output from T. The question is, how many of these do we need in order to determine that it is unlikely that U(k) has been generated randomly? If poor questioning technique or leading questions are employed, then U(k) would have to be vast. Given that U(k) is vast, how many sentences in C(k) will we need to ensure that we know when a valid sentence in U(k) is uttered?
And here is the crux of the matter: given that we don't know U(k), we cannot guarantee that C(k) is a reliable and indicative subset of U(k). If C(k) is unreliable, then none of the sentences in U(k) can be validated unless they can be separately validated through a reliable external mechanism. For those who don't know, heresay evidence doesn't really count here and, unfortunately, heresay evidence is used extensively to provide C(k). More importantly, if C(k) is used as a source for questions applied during the interrogation and there exists a set P(k), which are legitimate permutations of C(k) that exist within U(k), then an unknowledgeable participant can provide a credible superset of C(k) that appears to enhance the interrogator's knowledge of U(k).
And then this randomly generated knowledge becomes a fact, for the purposes of the interrogators, and may be used against the same person again, or on other people. It wouldn't surprise if real criminals have actually been let go because they had no knowledge of the largely fabricated scenarios that are generated by other people who have been pulled into the net.
Wednesday, June 18, 2008
Subscribe to:
Posts (Atom)