Fun in parts, but not for the characters as I recall. Unless there's a happy-ending version where the Savage and Lenina drive off into the distance under a voice-over from our Ford (i.e. Harrison).
IIRC, Lenina dies happily after finally escaping the hell from which she has been accidentally exiled (i.e., the Reservation). John (the Savage), however, dies because he completely fails to understand that other viewpoints than his own might be valid: that is, he's guilty of exactly the same sin that he accuses the World State of. (In fact, the World State does accept John's views as having merit: that's why they have the Reservation, and the Islands.)
Of course, John got a bum deal because of his upbringing... he's a child of both the World State and the Reservation and doesn't fit in with either. The World State would tolerate him, though, if he'd only given it a chance.
That's fine, but I'd put "error:" after the return success.
You didn't read what I said, did you? I can't put it at the end because you can't jump over a variable declaration in C++. (Because you can't jump into a greater scope than the one you're currently in, but you can jump into a lesser one.)
And I can't use exceptions because this is an embedded device where I can't afford the 2kB-per-stack-frame hit that enabling exceptions in gcc gives me.
I've found myself, somewhat to my horror, using the following idiom in C++ quite a lot:
function() {
if (false)
{
error:
printf("barf!\n");
return failure;
}
Object1 o1();
if (o.failed())
goto error;
Object2 o2();
if (o.failed())
goto error;...etc, several times...
return success; }
The jump target goes at the beginning, rather than the end, because C++ won't let you jump over a constructor. Previously, I was using nested scopes, which led to messy, overly indented code. This way at least produces cleaner code, although that 'if (false)' is a little startling... maybe I should consider wrapping it in a macro for readability's sake.
But seriously: whom would you prefer to do business with - the guy who constantly tinkers with his PSP, or the other guy who actually bothered to learn your language?
Given that anyone who learned the language wouldn't be interested in getting the translator software, I think a better question is:
Who would you prefer to do business with - the guy who constantly tinkers with his PSP and is making an effort to communicate, or the other guy who thinks you're mentally subnormal because you don't speak English?
I'd love to see the BBC's Jeremy Paxman interview George Bush, for example...
*splork*
I'd pay money to see that...
"So, Mr. Bush. They say you're an idiot. Are you?"
Re:I think I buy into this "ajax" thing
on
Ajax in Action
·
· Score: 1
...it's more about the architectural view of the application, where you think of the browser hosting your application rather than content and the server produces data rather than content...
Wait... this is like the MVC thing that Smalltalk invented 25 years ago, right? Where you divide your application up into the Model (the thing that manages the data), the View (the thing that takes the data from the Model and organises it for the user interface), and the Controller (the thing that the user actually interacts with)?
The only real difference I can see between AJAX and conventional web apps is that with a conventional app, the big network pipe is between the View and the Controller; with an AJAX app, the network pipe's between the Model and the View. Plus a small matter of coding, of course.
Every time I come across a cool new technology, it turns out that Smalltalk invented it in 1980...
Does anybody know whether we could use asteroids to transport probes around space?
Wouldn't an asteroid potential provide a fast and free transportation system? Wouldn't an asteroid potential provide a fast and free transportation system?
No, not really... because in order to get to the asteroid, you've got to match orbits with it, which means that you're already going to whereever the asteroid is going.
Wouldn't they provide rudmentary protection against space radaition somewhat?
This might be more useful, but given that the only major cause of radiation in space is the sun, and shielding machines against solar radiation isn't actually that hard, it's probably not worth it.
If you ask me, NASA and other space agencies should be firing out probes like crazy.
Absolutely.
Given that we now have two designs of Mars probe that have proven to be wildly successful --- the Spirit and Opportunity rovers, and the Pathfinder lander --- I think that the right think to do now is to mass produce them, bulk launch them, and drop a dozen of each on Mars. Because the design's been finalised, the cost per probe should be a fraction of what it was for the original; just work from the blueprints. And because you're launching lots, you can stand a higher failure rate, so you can target more interesting parts of the planet. That way you should be able to get some seriously interesting science, very cheaply.
(The reason why all the Mars probes are showing really boring bits of desert is because they've been deliberately targeted at wide, flat areas to increase the chances that they'll get down safely.)
If you want simple, static content, thttpd is stupidly tiny, stupidly scalable, and way faster than Apache. Unfortunately it uses the old fork model for dealing with CGI scripts which make it quite slow as doing that (but no worse than the old NCSA httpd). It has a number of interesting features, such as per-filetype bandwidth throttling (so you can specify that MP3 files only get transferred at 10kB/sec), but also has some suprising omissions --- the MIME type database is hard-coded, and it only handles HTTP 1.1. But if you have a simple site based mainly around static pages, thttpd is probably ideal for your purposes.
That falls down when trying to access my own websites from the LAN.
Shrug. Works for me --- I can access my web server (which is listening on all IP addresses) from the LAN using either name.
Using a laptop at home, then taking it elsewhere without changing settings like the NTP server, mailserver, etc.
Is that safe? What you're doing, in effect, is taking a machine that is connected to your secured LAN and plugging it in to the unsecured internet (and vice versa). There are entirely different security levels involved in the machine's two roles, and unless the machine is set up to have two completely different personalities, then you may have problems.
Wouldn't a better option be to use the machine only in its LAN personality, but when you're not at home connect it to your LAN via a secured VPN? Apart from anything else, a lot of the config issues should go away.
IPv6: it's autoconfigured. IPv4: you do have the DHCP server setup correctly, right?
Yes, I have a DHCP server --- but I only use it for guest machines. All my hard-wired machines have static addresses, allocated manually, because I don't want any addresses to change whenever I change network cards (or replace a machine completely).
No, it's not. I have to run split zone DNS so that hosts on my LAN know that "foo.example.com" is at 10.1.0.5, but that the rest of the Internet knows to connect to 123.45.67.8.
...which is doing things the complicated way; given that your local subnet and the internet are basically two different networks, with a gateway, don't try to force the same names. My main server is called 'pyanfar.local' (10.0.0.1) and 'gate.cowlark.com' (IP address I forget). No problem.
...as long as you don't want to use any P2P apps, VOIP, VPNs, or anything else that depends on a transparent end-to-end connection.
This is the only thing you mention that might actually be desirable; but I still don't want it. Firstly, most of the apps you mention work fine through NATs. Secondly, anything that doesn't, I want to know about so I can explicitely tunnel it through.
NAT is a horrible hack that provides no security above a default-deny firewall.
...which is why you have a real firewall as well (which you still need on IPv6).
Pv6 gives you globally-routable LANs and much less administrative overhead. That is why you want it.
Except I don't want a globally-routable LAN. I like having the inside and outside of my LAN two seperate worlds that operate by different rules. I much, much prefer having to route everything through the gateway, because it gives me a single device that handles all the security.
I'm sorry, but I simply don't want any of the features you've mentioned so far. Some of them are things I explicitly won't want --- dynamically allocated IP addresses, for example. (I have yet to get a straight answer as to whether IPv6 supports static IP addresses...)
Because, basically, IPv4 does everything I want it to do right now. I understand it. It's simple to set up. It's simple to maintain. I don't need additional IP addresses on my NATted system, because I can multiplex all my servers onto a single IP address using firewall rules (if I want to; I currently only have one server).
From what I've seen, IPv6 is hideously complicated, the stack is vast, there are innumerable incompatibility problems with various applications, and basically it's just too much effort for too little gain.
You won't run Linux on it, because it's a 16-bit 8086 clone inside with 256kB RAM and 512kB flash, but if you want something to provide a quick-and-easy ethernet interface to a piece of development hardware, it's ideal. I suspect it may be a bit expensive, though (no prices available).
There are some other similar devices from the same manufacturer --- among them is the WiPort, which is much the same thing as the XPort but uses 802.11b...
The brief summary of his idea seems to be that he's managing to con hydrogen atoms into giving up some of their energy to ionise helium. The end result is positive helium ions (which you siphon off and use to collect the energy), and 'hydrinos' --- hydrogen atoms at a lower energy state than classical theory predicts is possible.
What does a hydrino look like? How does it behave? An element's chemical properties are intrinsically tied up in its electron shell; and a hydrino has an electron shell that's significantly different from a conventional hydrogen atom. So, what chemical properties does a hydrino gas have?
This sort of thing is quite important. Deuterium is an isotope of hydrogen with an atomic mass of 2 instead of 1. It has noticeably different chemical properties to hydrogen, to such an extent that heavy water (water made with deuterium instead of hydrogen) is considered toxic.
If he's going to be producing large amounts of what amounts to a new element not found in nature and releasing it into the atmosphere, I'd expect him to do some careful chemical analysis of its properties first... nuclear physicists tend to be a bit careless when it comes to chemical properties of the atoms they're dealing with.
(Of course, because hydrinos don't seem to be found in nature despite having a lower energy level than ordinary hydrogen, I suspect he could well be talking out of his hat.)
As you say, fail-over would take place very quickly and most users wouldn't even notice anything strange. Someone would find a bunch of sites are suddenly inaccessible and e-mail is not functioning well.
Any business that uses email, or a VPN, or VoIP is going to find their productivity grinding to a halt because they rely on computerised communications too much. This includes hospital internal communications, government offices including things like welfare, banks, emergency services, all media, transportation... backup communications basically consist of the telephone; except big chunks of the phone system will start flaking out, because some telcos route stuff via the 'net (although they really shouldn't). What's left will be pretty much congested.
Any deaths that occur are not going to happen because fire-breathing robotic buses will start hunting down the hapless populace. They'll happen because some doctor didn't get hold of the medical files in time, or because a social security office was unable to check up on some old woman who hasn't been in to pick up her welfare cheque, or because a fire engine was unable to get through the traffic jam caused by the smart timing of a set of traffic lights going haywire because the computer that runs them can't find the traffic sensors any more, etc. The big stuff, banks, the military, the government, they'll all have working backup plans. The little stuff won't, because email and the web are always there.
There will also be more deaths further on down the line when the economic fallout hits. Imagine just how much money would be lost if every tech-related company in the country suddenly lost two-thirds of its productivity. For a lot of small businesses, that's enough to drive them under...
The US has no control over the internet, they can mess with it and poke it a little but nothing more.
You're kidding, right?
If the US government really wants to, they could shut down the root DNS servers, or even worse, set them to produce bogus data. That will cause, very quickly, worldwide chaos as the 'net becomes unusable. People will work around it very quickly --- I'm sure most clued-up governments have backup servers and all the major ISPs are set up to fail-over to them at the first sign of trouble --- but in the mean time, a hell of a lot of mission critical infrastructure will have gone belly up, all around the world. And then people really will die.
The US government would have to be idiots to do this because the US has as much mission-critical applications based on the 'net as everyone else (banks and such; forget the military, they've got their own networks). But... well... these days, the US government is not doing that great a job at persuading the world that they're not idiots, and that's what's making people nervous.
When there's a camera which can take pictures of the landing sites from earth, there will be pictures of the landing sites from earth. Not before.
Given that the landers took pictures on the way down and the way up, you could probably put together higher-resolution data for the areas around the landing sites; it should be possible to patch these in to the existing datasets, so you could do such things as examine the landing sites from the 'air' and get decent results.
If you like debian you'll love Gentoo. emerge, rc-update, etc-update can give you a fully up to date distro every day. The whole idea of major releases goes away.
I tried Gentoo once... thirty-six hours later, it was still compiling, so I gave up. Sorry, but binary packages are just far too convenient.
I love OpenBSD. I think it's an amazing OS for people who know what they're doing: secure, consistent, amazingly documented...
OTOH, I also love Debian. I think it's the best userland package management system I've ever seen. It's less flexible than BSD's roll-your-own userland, but far easier to manage.
What I'd really like is to be able to run Debian on top of the OpenBSD kernel. That would give me the best of both worlds: OpenBSD's stellar kernel, and Debian's amazing package management. I know there used to be projects underway to port Debian to FreeBSD and NetBSD, but they seem to have died, and frankly I'd rather use OpenBSD.
Slashdot Mirror
Oops --- you're right. It's been a long time since I read it... what happens to Lenina eventually?
IIRC, Lenina dies happily after finally escaping the hell from which she has been accidentally exiled (i.e., the Reservation). John (the Savage), however, dies because he completely fails to understand that other viewpoints than his own might be valid: that is, he's guilty of exactly the same sin that he accuses the World State of. (In fact, the World State does accept John's views as having merit: that's why they have the Reservation, and the Islands.)
Of course, John got a bum deal because of his upbringing... he's a child of both the World State and the Reservation and doesn't fit in with either. The World State would tolerate him, though, if he'd only given it a chance.
Isn't that a bit high-brow for Slashdot?
</snobbery>
Nice one. Must reread that sometime... it's a fun book (unlike 1984 which is miserable and harrowing and really unpleasant).
You didn't read what I said, did you? I can't put it at the end because you can't jump over a variable declaration in C++. (Because you can't jump into a greater scope than the one you're currently in, but you can jump into a lesser one.)
And I can't use exceptions because this is an embedded device where I can't afford the 2kB-per-stack-frame hit that enabling exceptions in gcc gives me.
I've found myself, somewhat to my horror, using the following idiom in C++ quite a lot:
...etc, several times...
function()
{
if (false)
{
error:
printf("barf!\n");
return failure;
}
Object1 o1();
if (o.failed())
goto error;
Object2 o2();
if (o.failed())
goto error;
return success;
}
The jump target goes at the beginning, rather than the end, because C++ won't let you jump over a constructor. Previously, I was using nested scopes, which led to messy, overly indented code. This way at least produces cleaner code, although that 'if (false)' is a little startling... maybe I should consider wrapping it in a macro for readability's sake.
Given that anyone who learned the language wouldn't be interested in getting the translator software, I think a better question is:
Who would you prefer to do business with - the guy who constantly tinkers with his PSP and is making an effort to communicate, or the other guy who thinks you're mentally subnormal because you don't speak English?
Specially designed for people who can't find their dick with both hands.
In fact, given that fact that only females have eggs, I think we can say for certain that it'll be her daughter...
Why, yes, I am incredibly pedantic. Why do you ask?
*splork*
I'd pay money to see that...
"So, Mr. Bush. They say you're an idiot. Are you?"
Wait... this is like the MVC thing that Smalltalk invented 25 years ago, right? Where you divide your application up into the Model (the thing that manages the data), the View (the thing that takes the data from the Model and organises it for the user interface), and the Controller (the thing that the user actually interacts with)?
The only real difference I can see between AJAX and conventional web apps is that with a conventional app, the big network pipe is between the View and the Controller; with an AJAX app, the network pipe's between the Model and the View. Plus a small matter of coding, of course.
Every time I come across a cool new technology, it turns out that Smalltalk invented it in 1980...
No, not really... because in order to get to the asteroid, you've got to match orbits with it, which means that you're already going to whereever the asteroid is going.
Wouldn't they provide rudmentary protection against space radaition somewhat?
This might be more useful, but given that the only major cause of radiation in space is the sun, and shielding machines against solar radiation isn't actually that hard, it's probably not worth it.
If you ask me, NASA and other space agencies should be firing out probes like crazy.
Absolutely.
Given that we now have two designs of Mars probe that have proven to be wildly successful --- the Spirit and Opportunity rovers, and the Pathfinder lander --- I think that the right think to do now is to mass produce them, bulk launch them, and drop a dozen of each on Mars. Because the design's been finalised, the cost per probe should be a fraction of what it was for the original; just work from the blueprints. And because you're launching lots, you can stand a higher failure rate, so you can target more interesting parts of the planet. That way you should be able to get some seriously interesting science, very cheaply.
(The reason why all the Mars probes are showing really boring bits of desert is because they've been deliberately targeted at wide, flat areas to increase the chances that they'll get down safely.)
...given the story title I was hoping for something a little bit more, well, ballistic.
...Irregular Webcomic?
If you want simple, static content, thttpd is stupidly tiny, stupidly scalable, and way faster than Apache. Unfortunately it uses the old fork model for dealing with CGI scripts which make it quite slow as doing that (but no worse than the old NCSA httpd). It has a number of interesting features, such as per-filetype bandwidth throttling (so you can specify that MP3 files only get transferred at 10kB/sec), but also has some suprising omissions --- the MIME type database is hard-coded, and it only handles HTTP 1.1. But if you have a simple site based mainly around static pages, thttpd is probably ideal for your purposes.
Shrug. Works for me --- I can access my web server (which is listening on all IP addresses) from the LAN using either name.
Using a laptop at home, then taking it elsewhere without changing settings like the NTP server, mailserver, etc.
Is that safe? What you're doing, in effect, is taking a machine that is connected to your secured LAN and plugging it in to the unsecured internet (and vice versa). There are entirely different security levels involved in the machine's two roles, and unless the machine is set up to have two completely different personalities, then you may have problems.
Wouldn't a better option be to use the machine only in its LAN personality, but when you're not at home connect it to your LAN via a secured VPN? Apart from anything else, a lot of the config issues should go away.
Yes, I have a DHCP server --- but I only use it for guest machines. All my hard-wired machines have static addresses, allocated manually, because I don't want any addresses to change whenever I change network cards (or replace a machine completely).
No, it's not. I have to run split zone DNS so that hosts on my LAN know that "foo.example.com" is at 10.1.0.5, but that the rest of the Internet knows to connect to 123.45.67.8.
...which is doing things the complicated way; given that your local subnet and the internet are basically two different networks, with a gateway, don't try to force the same names. My main server is called 'pyanfar.local' (10.0.0.1) and 'gate.cowlark.com' (IP address I forget). No problem.
This is the only thing you mention that might actually be desirable; but I still don't want it. Firstly, most of the apps you mention work fine through NATs. Secondly, anything that doesn't, I want to know about so I can explicitely tunnel it through.
NAT is a horrible hack that provides no security above a default-deny firewall.
...which is why you have a real firewall as well (which you still need on IPv6).
Pv6 gives you globally-routable LANs and much less administrative overhead. That is why you want it.
Except I don't want a globally-routable LAN. I like having the inside and outside of my LAN two seperate worlds that operate by different rules. I much, much prefer having to route everything through the gateway, because it gives me a single device that handles all the security.
I'm sorry, but I simply don't want any of the features you've mentioned so far. Some of them are things I explicitly won't want --- dynamically allocated IP addresses, for example. (I have yet to get a straight answer as to whether IPv6 supports static IP addresses...)
I'm not sold.
Because, basically, IPv4 does everything I want it to do right now. I understand it. It's simple to set up. It's simple to maintain. I don't need additional IP addresses on my NATted system, because I can multiplex all my servers onto a single IP address using firewall rules (if I want to; I currently only have one server).
From what I've seen, IPv6 is hideously complicated, the stack is vast, there are innumerable incompatibility problems with various applications, and basically it's just too much effort for too little gain.
Can anyone convince me otherwise?
XPort product page with lots of marketing spiel
You won't run Linux on it, because it's a 16-bit 8086 clone inside with 256kB RAM and 512kB flash, but if you want something to provide a quick-and-easy ethernet interface to a piece of development hardware, it's ideal. I suspect it may be a bit expensive, though (no prices available).
There are some other similar devices from the same manufacturer --- among them is the WiPort, which is much the same thing as the XPort but uses 802.11b...
What does a hydrino look like? How does it behave? An element's chemical properties are intrinsically tied up in its electron shell; and a hydrino has an electron shell that's significantly different from a conventional hydrogen atom. So, what chemical properties does a hydrino gas have?
This sort of thing is quite important. Deuterium is an isotope of hydrogen with an atomic mass of 2 instead of 1. It has noticeably different chemical properties to hydrogen, to such an extent that heavy water (water made with deuterium instead of hydrogen) is considered toxic.
If he's going to be producing large amounts of what amounts to a new element not found in nature and releasing it into the atmosphere, I'd expect him to do some careful chemical analysis of its properties first... nuclear physicists tend to be a bit careless when it comes to chemical properties of the atoms they're dealing with.
(Of course, because hydrinos don't seem to be found in nature despite having a lower energy level than ordinary hydrogen, I suspect he could well be talking out of his hat.)
Any business that uses email, or a VPN, or VoIP is going to find their productivity grinding to a halt because they rely on computerised communications too much. This includes hospital internal communications, government offices including things like welfare, banks, emergency services, all media, transportation... backup communications basically consist of the telephone; except big chunks of the phone system will start flaking out, because some telcos route stuff via the 'net (although they really shouldn't). What's left will be pretty much congested.
Any deaths that occur are not going to happen because fire-breathing robotic buses will start hunting down the hapless populace. They'll happen because some doctor didn't get hold of the medical files in time, or because a social security office was unable to check up on some old woman who hasn't been in to pick up her welfare cheque, or because a fire engine was unable to get through the traffic jam caused by the smart timing of a set of traffic lights going haywire because the computer that runs them can't find the traffic sensors any more, etc. The big stuff, banks, the military, the government, they'll all have working backup plans. The little stuff won't, because email and the web are always there.
There will also be more deaths further on down the line when the economic fallout hits. Imagine just how much money would be lost if every tech-related company in the country suddenly lost two-thirds of its productivity. For a lot of small businesses, that's enough to drive them under...
You're kidding, right?
If the US government really wants to, they could shut down the root DNS servers, or even worse, set them to produce bogus data. That will cause, very quickly, worldwide chaos as the 'net becomes unusable. People will work around it very quickly --- I'm sure most clued-up governments have backup servers and all the major ISPs are set up to fail-over to them at the first sign of trouble --- but in the mean time, a hell of a lot of mission critical infrastructure will have gone belly up, all around the world. And then people really will die.
The US government would have to be idiots to do this because the US has as much mission-critical applications based on the 'net as everyone else (banks and such; forget the military, they've got their own networks). But... well... these days, the US government is not doing that great a job at persuading the world that they're not idiots, and that's what's making people nervous.
Given that the landers took pictures on the way down and the way up, you could probably put together higher-resolution data for the areas around the landing sites; it should be possible to patch these in to the existing datasets, so you could do such things as examine the landing sites from the 'air' and get decent results.
I tried Gentoo once... thirty-six hours later, it was still compiling, so I gave up. Sorry, but binary packages are just far too convenient.
OTOH, I also love Debian. I think it's the best userland package management system I've ever seen. It's less flexible than BSD's roll-your-own userland, but far easier to manage.
What I'd really like is to be able to run Debian on top of the OpenBSD kernel. That would give me the best of both worlds: OpenBSD's stellar kernel, and Debian's amazing package management. I know there used to be projects underway to port Debian to FreeBSD and NetBSD, but they seem to have died, and frankly I'd rather use OpenBSD.
Oh, well...
But I couldn't find one that fit my monitor.