You'd think that the logical response from the HFT guys would be 'okay, let's put in a rule such that if the stock price drops by X amount, hold off from trading for Y seconds in case it's a market glitch'. But from some reports, that's exactly what happened and the fact that some of the HFTs were missing from the market caused the others to panic (well, as much as computers can; calmly fire their worst-case rules I guess). 'Nobody's bought this stock at all in the last second - it must be a real turkey! Sellsellsell!'
What I don't get is why the urgency to trade. Surely the stock market should be considered a system with very noisy, lossy data, and the sensible approach to volatility, if one was architecting a global system, would be to wait, step back, smooth out the noise and look for the true signal underneath. Which as I said before, really works on a yearly cycle - outside of genuine crisis response, most company information is only generated annually. Some of the really important planetary data - global fish stocks, species extinctions - only gets generated every decade, if we're lucky.
In between, all the market has to feed on is gossip, hype and its own internal psychological weather. If you were trying to build a system to make honest, practical, long-term investment decisions, you wouldn't look at the news of the day at all, for anything short of a nuclear war.
But trading seems to be about the opposite: look FOR volatility and exploit it - amplify it, even - before the other guy does. Get yours, and let the devil take the rest.
Which would be fine if this were a videogame, but it's the economies of nations which these guys are speculating with. At some point it becomes like drunk deadbeat dad playing poker with mom's housekeeping money. It's not fun anymore, it's not productive, and it's certainly immoral, when it's not outright criminal.
More and more the markets seem decoupled from reality. Why is it so hyper-urgent for a trade to complete in milliseconds, even if it means selling at rock-bottom price? Isn't that just really dumb programming?
Imposing a global circuit breaker seems like one way of fixing it... but why is the trading so frenetic in the first place? Why this absolute pressure to trade nownowNOW?
These are real companies people are betting on. Companies have lives in the years to decades, and at best their profits are measured in quarters - and even that's far too short-term thinking compared to human society, the biosphere and the ecological damage our industrial activities are doing.
There just isn't any meaningful data that can be generated about the activities of corporations on the millisecond scale. Not really any on less than a yearly scale, if you think about it. The biggest news right now is the Deepwater Horizon oil spill, and what's the timeline for fixing that? Weeks to months.
What does society actually gain from ultra-fast gambling on the markets? Other than a cheap thrill and massively increased risk?
A NASA Space/Mars Colony if anything would help us live better with smaller carbon footprint here on Earth.
During the last 40 years NASA has spent lots of R&D money on high efficiency solar Panels, Fuel Cells, Water Recycling/Pufification, all technologies required to live lightly on the land, or in the very finite resources available to Astronauts in space or on the Moon/Mars.
Or alternatively, we could do exactly what Jimmy Carter suggested in 1979 -- for which integrity and foresight he got booed out of the White House --
Launch an Apollo-class initiative to directly create and deploy technologies designed to reduce our energy use and live sustainably right here in Earth's biosphere
Oh, but living within our means will magically work in spaaaaace yet it won't on Earth where it's much more doable?
Don't rely on spin-offs. If you really want technology to save the Earth, invest in technology to save the Earth. Create a National Doing The Right Thing Agency. If you spend money on space, you might get a few tiny spinoffs, but most of it will go on special-purpose hardware which is useless on Earth.
Here's what Carter tried to do, and for which he was considered a failure:
"Point one: I am tonight setting a clear goal for the energy policy of the United States. Beginning this moment, this nation will never use more foreign oil than we did in 1977 -- never. From now on, every new addition to our demand for energy will be met from our own production and our own conservation. The generation-long growth in our dependence on foreign oil will be stopped dead in its tracks right now and then reversed as we move through the 1980s, for I am tonight setting the further goal of cutting our dependence on foreign oil by one-half by the end of the next decade -- a saving of over 4-1/2 million barrels of imported oil per day.
Point two: To ensure that we meet these targets, I will use my presidential authority to set import quotas. I'm announcing tonight that for 1979 and 1980, I will forbid the entry into this country of one drop of foreign oil more than these goals allow. These quotas will ensure a reduction in imports even below the ambitious levels we set at the recent Tokyo summit.
Point three: To give us energy security, I am asking for the most massive peacetime commitment of funds and resources in our nation's history to develop America's own alternative sources of fuel -- from coal, from oil shale, from plant products for gasohol, from unconventional gas, from the sun. "
It is an American - and global - tragedy that he didn't get a second term.
No kidding. You don't think I know what "one way ticket" means? Even if the provided life support was only designed to keep me alive for a few days, I'd still jump at the opportunity.
Even if your journey was entirely inside a sealed ship where you only thought you were going to Mars?
There lies the seed of a SF reality show, I'm sure.
Ending the cosmic equivalent of having all of our eggs in one basket? We're one natural disaster away from complete annhilation of our race. I'd kinda like to have at least a few people offworld just in case.
Given that any space colonisation effort will result in a solar-system-wide economic and transport infrastructure, and that the colonies will not be self-sustaining until a long time in the future and will require constant resupply and transfer of personnel from Earth... and once the colonies do become self-sustaining they will almost certainly continue to participate in the solar system economy and travel and trade will continue...... then exactly what class of human-extinction-level events would such colonies guard against which aren't already effectively defended against on Earth?
Asteroids? The colonies will be more vulnerable than Earth, and Earth has survived them before.
Ecological breakdown? The colonies will be FAR more vulnerable to this than Earth, and any technology capable of sustaining long-term life in space can be far more cheaply deployed on Earth to protect and restore our biosphere.
Plague? Due to the system-wide travel/trade grid, that's going to get into the cislunar shuttles as easily as it gets into jetliners, unless you shut down the entire Sol economy.
War? The colonists are going to come from multiple national allegiances and bring all their conflicts with them to space. Any war is going to affect the whole system, not just Earth. Plus, the expansion into space, particularly if it involves nuclear thrusters, will put more potential weapons of mass destruction into more hands. Every asteroid miner is a possible city-killer.
I just don't see how space expansion helps protect us. If anything, it just raises the stakes.
Because Drupal is something you run on one website, and zillions of individual websites with hand-crafted Drupal installs does not make a Facebook killer.
Conceivably one could build a distributed social network on Drupal, in the same way you could build it on PHP, but Drupal itself is still stuck in the 'one web server, one database, a bunch of pages/resources and lots of web clients viewing it' mould.
I don't understand why they don't use e-mail lists (my preferred way), but then again I don't understand synchronous texting either.
Because managing email lists is a royal pain. You either have a private in-client group which you send to, in which case one person needs to be the mailing list manager, or you subscribe to an email list manager service, which are... not so easy to find nowadays - and then you need to remember your MLM password, etc, etc.
Or, you create a Facebook group which is public, automatically advertised to all your friends, they can join at their convenience, and management becomes a breeze.
The ease of use of Facebook is about 1000x over mailing lists, and it's guaranteed to be always up to date, and that's what catapults it from 'toy' into 'standard communication channel'.
The other thing is that the concept of'Third Place is also very important. Email is a private-to-private communication channel. Web is a public-broadcast medium. But Facebook is deliberately architected as a gigantic Third Place hitting the spot in between: semi-public, semi-private, and discoverable. The use of real names and the purposeful lack of privacy is a huge element in this: you don't need to know someone's handle, because this is a medium for finding people, and for having overhearable discussions, not for having private chats. The network itself automatically reaches out and brings people in with its reminders and prompts. This makes it a big 'front porch in the sky' which our online (and even offline) culture hasn't had yet, and that's why it took off.
Yes, please build an open replacement for this. Online third place is far too important to leave to a private provider. But understand what it is that Facebook is doing right.
JOINs just don't scale. You can distribute data across any number of nodes, but JOINing data which lives on separate computers is not gonna happen.
If that's the case, then surely we're Doing Something Really Wrong with our implementation of relational theory. Should we perhaps be looking at things like Extended Set Theory instead?
Relational - (and more specifically, SQL, which as Chris Date is at pains to tell every is NOT even a correct let alone good implementation of the relational model - but even Codd's original paper shows signs of this) - came out of a timesharing environment, where it was just assumed as a matter of course that you'd have very large datasets sitting on single giant machines. And out of that environment came the 'database server' concept we're familiar with today.
The problem is that in the age of the Web, more and more our data IS distributed among thousands if not millions of nodes. This is especially obvious in social networking - at the moment, we're very obviously Doing It Wrong by relying on huge centralised mega-stores like Facebook. We should be asking why this is, and how our tools can be improved to stop forcing us toward a centralised model when decentralised is what we clearly need.
It seems to me that in a massively decentralised model, it ought to be easy enough to have multi-dataset JOINs or some equivalent. We just have to implement a way of caching the results and transmitting only the changes, in a store-forward-publish model.
Ie, a decentralised JOIN (or any other dataset filter) shouldn't be an 'operation' which runs, connects to lots of datasources, transmits/receives huge amounts of information, processes it into a subset, hands it on and then disconnects, throwing away all that data. That model is still harking back to the old batch processing days of mainframes. Instead, a JOIN should be a sort of object: you create a new derived dataset entity which connects ONCE to all its sources, grabs what it needs, then CACHES its data locally so it doesn't consume bandwidth.
This would let us create a truly decentralised 'web of data' - a bit like the old Usenet model, but for data not next - where you subscribe to datasets, receive just the updates, do whatever filtering or processing you want on the updates as they come in, and present your derived view - a bit like a SQL 'View' but far more general and able to do arbitrary Turing-complete processing.
I dunno who's working on this but it seems obvious that we need something like that. Huge centralised 'databases' are just so foreign to the Web model that I don't know why we still force people to use them.
This is the important point people don't seem to realise. It's not how much royalties cost, it's that you cannot legally require ANY royalties at all and be GPL compliant.
This is why Novell raised eyebrows with their Microsoft patent deal, remember? Because it's not legal.
If you want to kick the H.264 patent problem to the OS to keep it out of Firefox, fine - but now you've just caused a new problem in Linux GPL codebases, because it won't be legal to install H.264 there.
H.264 is only legal as non Open Source software. Which makes it a right nuisance, like binary driver blobs. It's not a good supportable way to go forward.
Nope. The man who has a watch always knows the time. A man who has two is never sure.
And the man who has three or four watches looks at the majority and is pretty sure again. There's a reason the Space Shuttle has more than two flight computers.
Slashdot Mirror
You'd think that the logical response from the HFT guys would be 'okay, let's put in a rule such that if the stock price drops by X amount, hold off from trading for Y seconds in case it's a market glitch'. But from some reports, that's exactly what happened and the fact that some of the HFTs were missing from the market caused the others to panic (well, as much as computers can; calmly fire their worst-case rules I guess). 'Nobody's bought this stock at all in the last second - it must be a real turkey! Sellsellsell!'
What I don't get is why the urgency to trade. Surely the stock market should be considered a system with very noisy, lossy data, and the sensible approach to volatility, if one was architecting a global system, would be to wait, step back, smooth out the noise and look for the true signal underneath. Which as I said before, really works on a yearly cycle - outside of genuine crisis response, most company information is only generated annually. Some of the really important planetary data - global fish stocks, species extinctions - only gets generated every decade, if we're lucky.
In between, all the market has to feed on is gossip, hype and its own internal psychological weather. If you were trying to build a system to make honest, practical, long-term investment decisions, you wouldn't look at the news of the day at all, for anything short of a nuclear war.
But trading seems to be about the opposite: look FOR volatility and exploit it - amplify it, even - before the other guy does. Get yours, and let the devil take the rest.
Which would be fine if this were a videogame, but it's the economies of nations which these guys are speculating with. At some point it becomes like drunk deadbeat dad playing poker with mom's housekeeping money. It's not fun anymore, it's not productive, and it's certainly immoral, when it's not outright criminal.
More and more the markets seem decoupled from reality. Why is it so hyper-urgent for a trade to complete in milliseconds, even if it means selling at rock-bottom price? Isn't that just really dumb programming?
Imposing a global circuit breaker seems like one way of fixing it... but why is the trading so frenetic in the first place? Why this absolute pressure to trade nownowNOW?
These are real companies people are betting on. Companies have lives in the years to decades, and at best their profits are measured in quarters - and even that's far too short-term thinking compared to human society, the biosphere and the ecological damage our industrial activities are doing.
There just isn't any meaningful data that can be generated about the activities of corporations on the millisecond scale. Not really any on less than a yearly scale, if you think about it. The biggest news right now is the Deepwater Horizon oil spill, and what's the timeline for fixing that? Weeks to months.
What does society actually gain from ultra-fast gambling on the markets? Other than a cheap thrill and massively increased risk?
you'd never call Captain Picard a Jedi
I find your lack of tea, Earl Grey, disturbing.
So, no, boiling water and turning it into a gas won't open a portal to a parallel universe.
No? You haven't seen my attempts at cooking.
Let's just say I keep a crowbar by the microwave.
You've got it backwards,
A NASA Space/Mars Colony if anything would help us live better with smaller carbon footprint here on Earth.
During the last 40 years NASA has spent lots of R&D money on high efficiency solar Panels, Fuel Cells, Water Recycling/Pufification, all technologies required to live lightly on the land, or in the very finite resources available to Astronauts in space or on the Moon/Mars.
Or alternatively, we could do exactly what Jimmy Carter suggested in 1979 -- for which integrity and foresight he got booed out of the White House --
Launch an Apollo-class initiative to directly create and deploy technologies designed to reduce our energy use and live sustainably right here in Earth's biosphere
Oh, but living within our means will magically work in spaaaaace yet it won't on Earth where it's much more doable?
Don't rely on spin-offs. If you really want technology to save the Earth, invest in technology to save the Earth. Create a National Doing The Right Thing Agency. If you spend money on space, you might get a few tiny spinoffs, but most of it will go on special-purpose hardware which is useless on Earth.
Here's what Carter tried to do, and for which he was considered a failure:
"Point one: I am tonight setting a clear goal for the energy policy of the United States. Beginning this moment, this nation will never use more foreign oil than we did in 1977 -- never. From now on, every new addition to our demand for energy will be met from our own production and our own conservation. The generation-long growth in our dependence on foreign oil will be stopped dead in its tracks right now and then reversed as we move through the 1980s, for I am tonight setting the further goal of cutting our dependence on foreign oil by one-half by the end of the next decade -- a saving of over 4-1/2 million barrels of imported oil per day.
Point two: To ensure that we meet these targets, I will use my presidential authority to set import quotas. I'm announcing tonight that for 1979 and 1980, I will forbid the entry into this country of one drop of foreign oil more than these goals allow. These quotas will ensure a reduction in imports even below the ambitious levels we set at the recent Tokyo summit.
Point three: To give us energy security, I am asking for the most massive peacetime commitment of funds and resources in our nation's history to develop America's own alternative sources of fuel -- from coal, from oil shale, from plant products for gasohol, from unconventional gas, from the sun. "
It is an American - and global - tragedy that he didn't get a second term.
No kidding. You don't think I know what "one way ticket" means? Even if the provided life support was only designed to keep me alive for a few days, I'd still jump at the opportunity.
Even if your journey was entirely inside a sealed ship where you only thought you were going to Mars?
There lies the seed of a SF reality show, I'm sure.
We are a dying civilization. The numbers prove it:
Thank you, Netcraft!
after WWII, after Disco.
And we barely survived WWII, but Disco.... brrrrr. We're still paying the price for that one.
Ending the cosmic equivalent of having all of our eggs in one basket? We're one natural disaster away from complete annhilation of our race. I'd kinda like to have at least a few people offworld just in case.
Given that any space colonisation effort will result in a solar-system-wide economic and transport infrastructure, and that the colonies will not be self-sustaining until a long time in the future and will require constant resupply and transfer of personnel from Earth... and once the colonies do become self-sustaining they will almost certainly continue to participate in the solar system economy and travel and trade will continue... ... then exactly what class of human-extinction-level events would such colonies guard against which aren't already effectively defended against on Earth?
Asteroids? The colonies will be more vulnerable than Earth, and Earth has survived them before.
Ecological breakdown? The colonies will be FAR more vulnerable to this than Earth, and any technology capable of sustaining long-term life in space can be far more cheaply deployed on Earth to protect and restore our biosphere.
Plague? Due to the system-wide travel/trade grid, that's going to get into the cislunar shuttles as easily as it gets into jetliners, unless you shut down the entire Sol economy.
War? The colonists are going to come from multiple national allegiances and bring all their conflicts with them to space. Any war is going to affect the whole system, not just Earth. Plus, the expansion into space, particularly if it involves nuclear thrusters, will put more potential weapons of mass destruction into more hands. Every asteroid miner is a possible city-killer.
I just don't see how space expansion helps protect us. If anything, it just raises the stakes.
And that's how come Jeff Goldblum was able to crash the mothership with a PowerBook.
There's no denying it, C is the basis of everything in computing.
If you consider 'everything in computing' to have begun on 1 January 1970, sure.
But strange as it is to contemplate now, there was life before Unix.
Now, if you said 'is the basis of every buffer overrun 0-day security flaw in computing...' then I'd have to agree.
Everyone who agrees with parent, please read this article NOW.
You don't learn C for the syntax, you learn it for the side effects.
And you learn Haskell to avoid the side effects...
panda bears that are large enough to shoot
But after eating and before leaving?
Specifically, Eric S Raymond.
John Sheridan would be proud.
Because Drupal is something you run on one website, and zillions of individual websites with hand-crafted Drupal installs does not make a Facebook killer.
Conceivably one could build a distributed social network on Drupal, in the same way you could build it on PHP, but Drupal itself is still stuck in the 'one web server, one database, a bunch of pages/resources and lots of web clients viewing it' mould.
I don't understand why they don't use e-mail lists (my preferred way), but then again I don't understand synchronous texting either.
Because managing email lists is a royal pain. You either have a private in-client group which you send to, in which case one person needs to be the mailing list manager, or you subscribe to an email list manager service, which are... not so easy to find nowadays - and then you need to remember your MLM password, etc, etc.
Or, you create a Facebook group which is public, automatically advertised to all your friends, they can join at their convenience, and management becomes a breeze.
The ease of use of Facebook is about 1000x over mailing lists, and it's guaranteed to be always up to date, and that's what catapults it from 'toy' into 'standard communication channel'.
The other thing is that the concept of'Third Place is also very important. Email is a private-to-private communication channel. Web is a public-broadcast medium. But Facebook is deliberately architected as a gigantic Third Place hitting the spot in between: semi-public, semi-private, and discoverable. The use of real names and the purposeful lack of privacy is a huge element in this: you don't need to know someone's handle, because this is a medium for finding people, and for having overhearable discussions, not for having private chats. The network itself automatically reaches out and brings people in with its reminders and prompts. This makes it a big 'front porch in the sky' which our online (and even offline) culture hasn't had yet, and that's why it took off.
Yes, please build an open replacement for this. Online third place is far too important to leave to a private provider. But understand what it is that Facebook is doing right.
I really don't get this. Everyone seems to be talking about Diaspora, which is still vaporware, when there are actual products that work right now.
There's an important pluralisation in that sentence. Products. And none of them talk to each other, right?
Sometimes more is less.
I think you'll find there's prior art in Otis Blackwell, 1957.
Course you can! You just need to be friends with the right cable guy and a few nanites.
And once you're done you not only get free TV, you get really cheap access to geosynchronous orbit.
JOINs just don't scale. You can distribute data across any number of nodes, but JOINing data which lives on separate computers is not gonna happen.
If that's the case, then surely we're Doing Something Really Wrong with our implementation of relational theory. Should we perhaps be looking at things like Extended Set Theory instead?
Relational - (and more specifically, SQL, which as Chris Date is at pains to tell every is NOT even a correct let alone good implementation of the relational model - but even Codd's original paper shows signs of this) - came out of a timesharing environment, where it was just assumed as a matter of course that you'd have very large datasets sitting on single giant machines. And out of that environment came the 'database server' concept we're familiar with today.
The problem is that in the age of the Web, more and more our data IS distributed among thousands if not millions of nodes. This is especially obvious in social networking - at the moment, we're very obviously Doing It Wrong by relying on huge centralised mega-stores like Facebook. We should be asking why this is, and how our tools can be improved to stop forcing us toward a centralised model when decentralised is what we clearly need.
It seems to me that in a massively decentralised model, it ought to be easy enough to have multi-dataset JOINs or some equivalent. We just have to implement a way of caching the results and transmitting only the changes, in a store-forward-publish model.
Ie, a decentralised JOIN (or any other dataset filter) shouldn't be an 'operation' which runs, connects to lots of datasources, transmits/receives huge amounts of information, processes it into a subset, hands it on and then disconnects, throwing away all that data. That model is still harking back to the old batch processing days of mainframes. Instead, a JOIN should be a sort of object: you create a new derived dataset entity which connects ONCE to all its sources, grabs what it needs, then CACHES its data locally so it doesn't consume bandwidth.
This would let us create a truly decentralised 'web of data' - a bit like the old Usenet model, but for data not next - where you subscribe to datasets, receive just the updates, do whatever filtering or processing you want on the updates as they come in, and present your derived view - a bit like a SQL 'View' but far more general and able to do arbitrary Turing-complete processing.
I dunno who's working on this but it seems obvious that we need something like that. Huge centralised 'databases' are just so foreign to the Web model that I don't know why we still force people to use them.
And Linux doesn't matter? Is this still Slashdot?
+++!
This is the important point people don't seem to realise. It's not how much royalties cost, it's that you cannot legally require ANY royalties at all and be GPL compliant.
This is why Novell raised eyebrows with their Microsoft patent deal, remember? Because it's not legal.
If you want to kick the H.264 patent problem to the OS to keep it out of Firefox, fine - but now you've just caused a new problem in Linux GPL codebases, because it won't be legal to install H.264 there.
H.264 is only legal as non Open Source software. Which makes it a right nuisance, like binary driver blobs. It's not a good supportable way to go forward.
Nope. The man who has a watch always knows the time. A man who has two is never sure.
And the man who has three or four watches looks at the majority and is pretty sure again. There's a reason the Space Shuttle has more than two flight computers.
H.264 is not going away. Those that make money from it can afford the license. For those that don't, it is free.
No, it is not free. It's available without paying money from certain manufacturers, but it's not free. It's jail time if you distribute it.
If it were free we wouldn't be having this conversation.