Well, there WAS a "warning" pheromone. You were supposed to spray it around dangerous places. It just meant "stay away". Since the game wasn't on branches but on ground, the "no entry" pheromone made less sense. Anyway, get the first colony somewhat running, just mark the first food supply, then change the profile of egg production so that all 3 classes are produced in equal amounts, not the default queens being just a small percent, then go to the macromanagement map and start spreading the queens all over the lawn. Some will die, some will survive and produce more queens. Really soon you will own the game.
Yep. Sometimes the code looks unoptimal, with some null padding and redundant jumps. You change it and things suddenly stop working, despite the fact the code runs ok in the debugger and all you did was some clean-up. But the code gets copied to some magical memory area. The null-padding is skipped and matches some magic values/registers that shouldn't be overwritten. Changing the length of the program breaks stuff. There is an unchecked race condition carefully avoided by counting CPU cycles (extreme optimization) and further optimizing one thread makes it trigger the race condition problem because it ends too early. You move a section of code and suddenly target of JMP is more than 127 bytes from its source. Things are ugly down there. And if program is self-modifying? If you perform maths on a jump address, to skip into a specific point of a just-generated matrix of procedures? double(**a)(); ?
Not all that new thing. They do. And remains of a dead ant work about the same way. So if you want them to forbid entry to some place, if there's an estabilished "road", stream of ants, too late, you need to use some mass destructiom chemical weapons. But if you see a single ant scout, smash it and smear it over the likely road, it will quite effectively stop other ants from, say, climbing the garden table legs.
In other cases running a handful of Windows boxes for the people that really need it mixed with a mostly Linux deployment is the answer.
One more option. Windows boxes with Open Source software running. If there are 8 things the box must do, and one of them is available only for Windows (find a good driver for a big CNC milling machine for Linux... and a CAD/CAM software to go with it), use free software wherever it fits. OOo instead of MS Office. Firefox, Thunderbird, Gimp, Inkscape... and if needed, Cygwin for most of the rest. Proprietary here: Corel Draw (sorry, Inkscape doesn't live up to my needs YET), IrfanView (well, nothing better Free ported to Windows), AVG (no Free antiviruses) and the CAD program plus the CNC driver. All the rest - free.
Because there are countless places where they may occur. In a strictly contained environment, when you get given number of fixed size data blocks, they are easy to eliminate. Whether they are eliminated or not, is a different matter. Just matter of will to eliminate them, not ability.
Re:Adblocking is nothing new...
on
KDE 3.5 Released
·
· Score: 1
That's not adblock(the extension) compatibile. Lots of false positives, no easy way of adding more rules, no site-based blocking, and above all, the format is different than adblock filterset (wildcards, regexp matching etc). Plus it doesn't stop the ads from loading, just from displaying.
I wonder if the adblocker from Konqueror is compatibile with firefox Adblock.
(...As you've certainly noticed...) Adblock by itself is worthless. Its empty filter base makes it inactive and only weeks of careful building it would make the extension normally useful. Only combined with a good killfile like Filterset.G it really kicks ass, at once. Same applies to any other adblocker - what filters are available for Konqueror?
And that's why it won here. This distro IS true Realtime OS, with kernel modified to work in realtime. In this it's equal with all the other available RTOSes, or even a bit below, because the support for realtime operation is young and not fully-featured. But while other RTOSes focused on adding more features, making it more stable and such, while neglecting actual efficiency, plain vanilla linux was developed to be a speed monster, with all that extra schedulers, optimizations, support for custom architectures, SMP and all the stuff that just made it a very FAST OS. Then the RealTime extension was added.
Make no mistake, Realtime doesn't mean fast. On normal systems, no matter how much CPU power, you can't guarantee some thread won't be stalled for longer than X miliseconds. On RTOS, given certain hardware speed and certain software load you can -guarantee- some threads will be given time within some fixed time. Often longer than they usually get on non-RTOS, but never longer than certain X miliseconds. Now thing is, how much can the X be, and what does it depend on? Well, certainly on amount of $$$ you put in the hardware, more CPU power = more spare CPU power that can be given sooner. So theoretically: Give enough CPU power, have arbitrarily short guaranted response time, down to time of one loop over the kernel procedures. In reality: You have just as much hardware, and the kernel of the RTOS eats up most of the resources, and due to all the failsafe checks, runs quite slowly in fact. True, at a constant safe pace that allows for granted 20 frames per second of input sensors analysis. But if you want 30, sorry, it starts crubmling, CPU overloaded, failsafes launched, frames lost. As long as you tell it to grab 20 frames a second, it won't fail, ever, no matter what though (as opposed to non-RTOS, which may lose a few frames just because it needed to swap out some memory or run a cron job in the background). All RTOSes do this. x MIPS, y RAM, m miliseconds for response, n miliseconds when the thread must finish or it will be forced to finish. Plus costs, reliablity and all the standard issues with any generic OS.
Now, given certain speed of hardware, what speed can you guarantee? Depends on the OS, and the faster the OS, the better the speed. And while Linux is really very fast, others aren't so. Writing a system that GUARANTEES 50 fps instead of 20 fps is damn hard. But writing a normal system that does 90fps on the average, without lowest speed guarantee? Well, possible, not so hard. Just keep it optimal. Then throw in the extra bits that make it a RTOS, and you notice that even after the RTOS overhead it never gets below 70fps. Check most pessimistic scenarios and you see it will never drop below 50fps.
If they shoot such a missile at my rocket, I may demand them to send me the source code. After all, they just distributed open source software. All over the area...
He's got a point though. It's just a tool, and should feel pretty "transparent" too. The important thing is the webpages, their content. How much time do you spend talking about your TV set as opposed to TV shows and movies? Or about your computer monitor? Without it, using the computer would be "somewhat" harder, but it just works, fine, let it be. You don't talk about brands of paper used for making your favourite magazine, about how amazing microphone did the singer on the concert use, or such. These are just media, and as such, the perfect ones are the ones that aren't visible at all, bringing the content directly. The more the medium is visible, the less the content it presents is, and it's all about the content.
But techniques of writing software immune to buffer overflows are known for decades. Your buffer is meant for 256 bytes of data, somebody sends a few K of data, your app simply keeps count of bytes streamed in and discards everything above the 255th byte. In case of predefined, fixed raw data blocks obtained from readout devices like radiotelescopes (as opposed to random user input), simple sanity checks allow for eliminating all buffer overflow risks.
The "malicious signal" may be of earth origin, just send it to the antennas on the right frequency and make it similar enough in shape to the space noise and it will get processed just the same. Or hijack a DNS and post new "work units" with malicious content acting as SETI. Thing is you don't need to separate the data, you just need to make the processing software secure, in such a way that data is analysed and never executed, there's no chance of buffer overflow or other potential risks coming from the data. Simple as that.
Never. 'Cause it never happens on such a scale, and I doubt anyone using these obscure languages would do it for other purposes than for fun. True, it may feel cool to translate Firefox to Esperanto, but how many people speaking Esperanto can't use Firefox in English version? Same here, so what if you translate Office to some obscure language 200 people on Earth speak? These 200 wouldn't get far if they didn't know some other, more common language. Most likely one which Office is already available in. I applaud this move loudly. Because it will cost Microsoft a fortune and might expand its market by less than 1%. And of course increase maintenance cost significantly (all patches and upgrades need to be translated to the new languages), increase development time and cost, defer development from bugfixes to some obscure features present only in some god-forgotten languages, and the profits will be marginal. So, all the way, Go Microsoft, translate more.
btw its a good idea to disable karma bonus if replying to trolls it lowers you maximum potential karma loss and reduces the chance of accidental mod-downs caused by re-parenting of your comment.
Not always. With some moderations like "troll", the post immediately loses "+1 Karma bonus", dropping to 0 from 2, and can be modded just -1 more. Plus, I can afford it:)
No, no, no. That won't pass. It's like a spyware "spyware detector" program that offers you to remove itself for a fee. "Get Firefox with Adblock, you won't be annoyed with ads like this one." This is just wrong.
As soon as I get a C64, I'll install Lunix on it, and then use it as a remote console. BTW, whenever you see a troll formatted to 60 columns width, it's very likely that it was posted from a Lunix box.
Slashdot Mirror
Well, just get a colony of ants in a lab to mass-produce it.
Forever, if the chemicals are produced from live(->dead) ants. probably easiest/cheapest source too.
Well, there WAS a "warning" pheromone. You were supposed to spray it around dangerous places. It just meant "stay away". Since the game wasn't on branches but on ground, the "no entry" pheromone made less sense.
Anyway, get the first colony somewhat running, just mark the first food supply, then change the profile of egg production so that all 3 classes are produced in equal amounts, not the default queens being just a small percent, then go to the macromanagement map and start spreading the queens all over the lawn. Some will die, some will survive and produce more queens. Really soon you will own the game.
If you have no clue about computers, Open Source is just as awful experience as flying if you have no wings.
As you prove, bad code does too.
For I Equals One
To MAX_FUNCTION_ITERATE
Do Function; Next I
Never hardcode numerical constants, even in Haikus!
Yep.
Sometimes the code looks unoptimal, with some null padding and redundant jumps. You change it and things suddenly stop working, despite the fact the code runs ok in the debugger and all you did was some clean-up.
But the code gets copied to some magical memory area. The null-padding is skipped and matches some magic values/registers that shouldn't be overwritten. Changing the length of the program breaks stuff. There is an unchecked race condition carefully avoided by counting CPU cycles (extreme optimization) and further optimizing one thread makes it trigger the race condition problem because it ends too early. You move a section of code and suddenly target of JMP is more than 127 bytes from its source.
Things are ugly down there.
And if program is self-modifying?
If you perform maths on a jump address, to skip into a specific point of a just-generated matrix of procedures?
double(**a)(); ?
DON'T TOUCH.
Not all that new thing. They do. And remains of a dead ant work about the same way. So if you want them to forbid entry to some place, if there's an estabilished "road", stream of ants, too late, you need to use some mass destructiom chemical weapons. But if you see a single ant scout, smash it and smear it over the likely road, it will quite effectively stop other ants from, say, climbing the garden table legs.
In other cases running a handful of Windows boxes for the people that really need it mixed with a mostly Linux deployment is the answer.
One more option. Windows boxes with Open Source software running. If there are 8 things the box must do, and one of them is available only for Windows (find a good driver for a big CNC milling machine for Linux... and a CAD/CAM software to go with it), use free software wherever it fits. OOo instead of MS Office. Firefox, Thunderbird, Gimp, Inkscape... and if needed, Cygwin for most of the rest.
Proprietary here: Corel Draw (sorry, Inkscape doesn't live up to my needs YET), IrfanView (well, nothing better Free ported to Windows), AVG (no Free antiviruses) and the CAD program plus the CNC driver. All the rest - free.
Well, the levels rise, but are they the source of the feeling or more like effect of it?
Tell me why the stars do shine
Tell me why the ivy twines
Tell me why the sky's so blue
and I'll tell you why I love you.
Nuclear fusion makes stars to shine
Phototropism makes ivy twine
Rayleigh scattering makes sky so blue
Sexual hormones are why I love you.
Because there are countless places where they may occur. In a strictly contained environment, when you get given number of fixed size data blocks, they are easy to eliminate.
Whether they are eliminated or not, is a different matter. Just matter of will to eliminate them, not ability.
That's not adblock(the extension) compatibile. Lots of false positives, no easy way of adding more rules, no site-based blocking, and above all, the format is different than adblock filterset (wildcards, regexp matching etc). Plus it doesn't stop the ads from loading, just from displaying.
I wonder if the adblocker from Konqueror is compatibile with firefox Adblock.
(...As you've certainly noticed...) Adblock by itself is worthless. Its empty filter base makes it inactive and only weeks of careful building it would make the extension normally useful. Only combined with a good killfile like Filterset.G it really kicks ass, at once. Same applies to any other adblocker - what filters are available for Konqueror?
but it wasn't built for RT from the ground up.
And that's why it won here.
This distro IS true Realtime OS, with kernel modified to work in realtime. In this it's equal with all the other available RTOSes, or even a bit below, because the support for realtime operation is young and not fully-featured. But while other RTOSes focused on adding more features, making it more stable and such, while neglecting actual efficiency, plain vanilla linux was developed to be a speed monster, with all that extra schedulers, optimizations, support for custom architectures, SMP and all the stuff that just made it a very FAST OS. Then the RealTime extension was added.
Make no mistake, Realtime doesn't mean fast. On normal systems, no matter how much CPU power, you can't guarantee some thread won't be stalled for longer than X miliseconds. On RTOS, given certain hardware speed and certain software load you can -guarantee- some threads will be given time within some fixed time. Often longer than they usually get on non-RTOS, but never longer than certain X miliseconds.
Now thing is, how much can the X be, and what does it depend on? Well, certainly on amount of $$$ you put in the hardware, more CPU power = more spare CPU power that can be given sooner. So theoretically: Give enough CPU power, have arbitrarily short guaranted response time, down to time of one loop over the kernel procedures. In reality: You have just as much hardware, and the kernel of the RTOS eats up most of the resources, and due to all the failsafe checks, runs quite slowly in fact. True, at a constant safe pace that allows for granted 20 frames per second of input sensors analysis. But if you want 30, sorry, it starts crubmling, CPU overloaded, failsafes launched, frames lost. As long as you tell it to grab 20 frames a second, it won't fail, ever, no matter what though (as opposed to non-RTOS, which may lose a few frames just because it needed to swap out some memory or run a cron job in the background).
All RTOSes do this. x MIPS, y RAM, m miliseconds for response, n miliseconds when the thread must finish or it will be forced to finish. Plus costs, reliablity and all the standard issues with any generic OS.
Now, given certain speed of hardware, what speed can you guarantee? Depends on the OS, and the faster the OS, the better the speed. And while Linux is really very fast, others aren't so. Writing a system that GUARANTEES 50 fps instead of 20 fps is damn hard. But writing a normal system that does 90fps on the average, without lowest speed guarantee? Well, possible, not so hard. Just keep it optimal. Then throw in the extra bits that make it a RTOS, and you notice that even after the RTOS overhead it never gets below 70fps. Check most pessimistic scenarios and you see it will never drop below 50fps.
unless we nuke or otherwise genocide ourselves first, that is.
If they shoot such a missile at my rocket, I may demand them to send me the source code. After all, they just distributed open source software. All over the area...
Too bad it doesn't run under Linux ;)
He's got a point though. It's just a tool, and should feel pretty "transparent" too. The important thing is the webpages, their content. How much time do you spend talking about your TV set as opposed to TV shows and movies? Or about your computer monitor? Without it, using the computer would be "somewhat" harder, but it just works, fine, let it be. You don't talk about brands of paper used for making your favourite magazine, about how amazing microphone did the singer on the concert use, or such. These are just media, and as such, the perfect ones are the ones that aren't visible at all, bringing the content directly. The more the medium is visible, the less the content it presents is, and it's all about the content.
But techniques of writing software immune to buffer overflows are known for decades. Your buffer is meant for 256 bytes of data, somebody sends a few K of data, your app simply keeps count of bytes streamed in and discards everything above the 255th byte. In case of predefined, fixed raw data blocks obtained from readout devices like radiotelescopes (as opposed to random user input), simple sanity checks allow for eliminating all buffer overflow risks.
The "malicious signal" may be of earth origin, just send it to the antennas on the right frequency and make it similar enough in shape to the space noise and it will get processed just the same. Or hijack a DNS and post new "work units" with malicious content acting as SETI.
Thing is you don't need to separate the data, you just need to make the processing software secure, in such a way that data is analysed and never executed, there's no chance of buffer overflow or other potential risks coming from the data. Simple as that.
Never. 'Cause it never happens on such a scale, and I doubt anyone using these obscure languages would do it for other purposes than for fun. True, it may feel cool to translate Firefox to Esperanto, but how many people speaking Esperanto can't use Firefox in English version?
Same here, so what if you translate Office to some obscure language 200 people on Earth speak? These 200 wouldn't get far if they didn't know some other, more common language. Most likely one which Office is already available in.
I applaud this move loudly. Because it will cost Microsoft a fortune and might expand its market by less than 1%. And of course increase maintenance cost significantly (all patches and upgrades need to be translated to the new languages), increase development time and cost, defer development from bugfixes to some obscure features present only in some god-forgotten languages, and the profits will be marginal.
So, all the way, Go Microsoft, translate more.
http://www.superdickery.com/
Superman is a dick.
(Now I'd LOVE to see a movie that contains a good compilation of events from this site.)
btw its a good idea to disable karma bonus if replying to trolls it lowers you maximum potential karma loss and reduces the chance of accidental mod-downs caused by re-parenting of your comment.
:)
Not always. With some moderations like "troll", the post immediately loses "+1 Karma bonus", dropping to 0 from 2, and can be modded just -1 more.
Plus, I can afford it
getting one early next year after the price drops.
and the modchip is released.
(that way they will actually cost Microsoft money...)
No, no, no. That won't pass. It's like a spyware "spyware detector" program that offers you to remove itself for a fee. "Get Firefox with Adblock, you won't be annoyed with ads like this one." This is just wrong.
As soon as I get a C64, I'll install Lunix on it, and then use it as a remote console. BTW, whenever you see a troll formatted to 60 columns width, it's very likely that it was posted from a Lunix box.
Unfortunately, no Firefox for Lunix. Sorry.