I'm an EE and I disagree. You're thinking of combinational logic, wiring together bunches of NAND and NOR gates as a heap of random logic. This leads to lots of fun with race conditions that introduce logic glitches.
Asynchronous logic uses a handshake signal to indicate completion of an operation. This is often done by adding an extra signal along with the result, the signal is asserted when the operation of the gate is completed.
There are other problems that have stopped asynchronous circuits from wide adoption. The handshake signalling adds overhead. More circuitry and more wiring, which means more silicon area used and higher wiring congestion.
It depends on the efficiency of the devices. A lightbulb isn't designed to radiate infrared energy, its designed to radiate energy in the visible light spectrum. Yes, infrared energy is radiated, but thats waste energy (which could be put to use).
A furnace is designed to be most efficient at radiating infrared energy. Other wavelengths are emitted but again they're just waste.
If you really believe that the amount of energy used by leaving on a lightbulb (or computer, or blender etc) continuously is offset by the furnace being operated on a lower duty cycle you're wrong. Just as wrong as if you insisted you don't need lights in the winter since the furnace is on and emitting visible light radiation.
If this ever becomes untrue I'll fully expect to live in an EZ-bake house.
Oregon, though with a significant technology industry, continues to use pencil and drawing board in most high school drafting classes. This does little to assist students in career preparation, whether they're interested in drawing PCBs or kitchen cabinets.
If you can find software that fits the bill thats great, however you're wrong in thinking that the lack of software does little to prepare a student for a career. Most of CAD is knowing how to design, not how to click buttons. In mechanical or architectural drafting there are a few students in a typical class who get it, they can make the transistion from hashed lines and bold lines to a three dimensional object. They can understand how things fit together and how the materials can interact. These are the skilled workers. In designing printed circuit boards there are a few students who understand signal integrity and how their choices will effect it.
The CAD tool part of it is easy to pick up isn't a fundamental part of the skills I consider important. When I'm interviewing candidates direct out of school my eyes glaze when they tell me how many software packages they know how to use or how many languages they've used. Tell me about a project you've been involved in, tell me how you went about the design project.
The tools I use now will be obsolete in a few years, sooner if I change companies. The skills I have will be valuable forever.
I consider knowing languages analogous to knowing how to flip burgers. Sure, you've flipped burgers at McDonald's, that doesn't mean you're a chef.
The evidence doesn't suggest anything. Music sales were up from last year. Well, music sales were up the year before that and the year before that and so on. You've got to go through some statistical evaluation of the numbers before you can say "Napster sold more records". Anything less is a lie.
The RIAA sold less CD singles last year, they also can't attribute that to Napster without going through the statistics to prove it or they're lying as well. Actually, they've already lied, Napster has lied and Napster supporters have lied.
Personally I'm pro mp3 sharing, moreso now that the RIAA has started bullying efforts. I'm not necessarily pro-Napster. It's a corporate entity trying to make money enabling other people to share music they have no copyright on. Get rid of the corporate pig in the middle and I fully support it. I think the record industry and electronic distribution can co-exist, but it will be a less powerful record industry.
Real time software mpeg2 encoding can't be done with all of the following three components at the present time: good picture quality, fast frame rate, good compression. The CODEC is just too math intensive for present processors. There are consumer MPEG2 encoding boards but as far as I know none of these have Linux support. The Hauppauge WinTV PVR is one such card, and retails for around $230.
If anybody can find inexpensive hardware MPEG2 encoders that have linux support I'd really appreciate it, as I've started playing with digital video and my DV camcorder.
The TiVO has hardware MPEG2 encoding and runs linux but I don't think you can get the drivers or hardware.
Hrm, just kind of a neat observation. I'm a hardware geek in a long line of artists. From talking with my father and sister I've realized that they see and remember differently than me. They see details that I don't see, but after they're described I can sort of see some of them. Things like the true colour of objects (heck, to me there's 8, just like in any respectable box of crayolas), the way light and shadow fade into each other and texture. They both seem to remember a great deal of these details as well.
I've improved myself somewhat by following through on part of Betty Edward's The New Drawing on the Right Side of the Brain but I still don't have the innate ability. It's still pretty interesting though, the sense of everything disappearing from my consiousness except for the artwork is very similar to what happens for me when working on a difficult and interesting engineering problem. I'm going to finish the book when I'm less overwhelmed with work.
First of all, ignore the "don't take either" comments. Learning a language is the smallest part of the skills a Computer Scientist or a Computer Engineer. New languages are introduced and old ones fall out of favour, you can pick new ones up fairly easily. Computer Science does deal mostly with software, but in a good program shouldn't deal mostly with programming languages. It should deal with algorithms, optimization and other things that help you write stable and maintainable code in any modern language. You'll be exposed to languages in the process, but the language shouldn't be the goal of the course, just a mechanism to get to the goal. Take courses you're interested in, if you like graphics take those courses. You'll still be learning the basics I mentioned however, just in a more digestible format for you.
Computer Engineering is more about the how's and why's of the guts and external interfaces of a computer system. How do architectural decisions effect performance and scaling for instance? How do different cache sizes and mechanisms effect performance for given applications? You'll spend time writing software and you'll spend time designing hardware and often you'll be doing both at the same time.
You can avoid university and still do well, some notable people have taken this route. Most become code monkeys however, they implement others work rather than architect the systems (software and/or hardware) that they work on.
I'm an electrical engineer with a lot of computer science courses, so I'm a bit of an outsider. I've seen the curriculum of my more (and less) succesful friends though, and the above is my take on the succesful programs.
Early adopters always pay the most. Every nVidia product is first released in an insanely expensive configuration up front. People with the available funds will purchase these, most other people won't.
Every nVidia product is next release in a feature reduced version, often available in white box OEM versions, for a fraction of the cost. Just like the GeForce 2 MX there will be a GeForce 3 MX.
The performance won't be as good, but it'll be available for a more reasonable price.
I don't care about banner ads as long as they don't get in the way of the reason I'm visiting a site: the content.
I no longer visit a number of sites because they display pop-up ads, JAVA (I assume anyway, they're interactive) ads which crash or slow my browser, or sites that display more advertisement than material. I could set things up so that the annoyances wouldn't appear, but I'd rather just ignore the site. Oh yeah, add rapidly flashing ads to the list of things that will turn me off (i.e., the morons at X-10)
All of this can be distilled down to an observation that advertisers should listen to. No matter how slick you think the advertisement is, no matter how many hours you spent on it or no matter how "ground breaking" it is, you've failed miserably if you turn off your target audience.
Market droids should also realize that it isn't the size of their adds that are stopping me from clicking through, its the fact that I already know the product or service or I'm just not interested.
The definition depends on the scope, in the scope of your argument your definition is insufficient. It may be correct as far as a dictionary entry goes, but is insufficient as far as it pertains to the United States patent law.
United States patent law includes provisions that put patented inventions in the public domain after a period of time has elapsed. This provides freedom for people to build on, extend and improve on the patented entity.
If you're going to play dictionary definition rebuttal then I'll propose that the GPL is by definition... not free.
freedom (frdm)
n.
1.The condition of being free of restraints.
The GPL restrains me from taking the linux kernel and marketing it as the closed source product eXtro's K-rad Kernel.
I don't agree that I should be able to do this, but it is an example of your practice of using the dictionary as a bludgeon to quell observations you don't agree with. In this case it "proves" the untrue assertion that the GPL is anti-freedom.
I can't imaging risking my life and my family's happiness for $600k + a promise of an
additional $800K ofer 15 years. That's just $100K a year, assuming that the Russian will
eventually pay the additional $800k, to commit treason, which is punishable by DEATH!
This guy was already making $110k/year salary, so this only doubled his salary.
If I double your salary, could I take a couple pot shots at you with my 9mm? What was
this guy on?!
I agree, but then people risk their lives and happiness for a lot less money than that. I'm not talking espionage, but what about the police, firefighters and armed forces personnel? Perhaps you could argue that they feel they're doing something for the good of all and that is the real reward.
What about coal miners? The point is that there are a lot of people doing dangerous jobs for not very much money. We may never know his motivation entirely, until they start to delve into his personal affairs. Gambling debts? Mistresses? Illegitimate children? Money probably wasn't the real factor, more likely it was a needed influx of money to bury some other problem.
Just conjecture, but from everything I've read recruiting for espionage (whether corporate or national) is usually rooted in finding some personal problem and exploiting it with the proper carrot.
There isn't a cut and dry answer in my opinion. It all depends on the data you're trying to destroy. The more undesirable the outcome if the data escapes the greater the lengths you've got to go to obliterate it.
If you're a newlywed and getting rid of your vast pornography collection a format is probably sufficient. Rewrite over the data with a random pattern if you're paranoid your new wife will use Norton Marriage-Breaker 2K1.
If you're dealing in contraban, whether the contraban is on the harddrive or the harddrive just contains incriminating evidence, destruction of the device would be your best bet. Remove the platters, stick them in the microwave for a few seconds (mostly just for the fireworks), sand them, sand blast them and melt them.
Just like cryptography the lengths you should go to depend on the difficulty in going to those lengths, the probability that those lengths can be compromised and the cost if those lengths are compromised.
Those are two seperate issues. The goal of copywrite is to protect the author and bolster creativity. It's not at all to protect the authors family.
With or without copywrite Grant still could've written his memoirs and expected them to provide for his family. They've got something, his memoirs, which others don't have. It's up to them (or him prehumously) to negotiate favourable terms.
"promote the Progress of Science and the useful Arts" and it only allows Congress to do that by
"securing for limited Times to Authors and Inventors the exclusive Right to their respective
Writings and Discoveries."
Which would indicate that copyright beyond the life of the author would be unjust, let alone an extension. (I realize you're disagreeing, but I'm amplifying a point) So the present copyright law, which states that a material is copyright for the life of the author + X years, is already unconstitutional.
The copyright law wasn't intended to protect or compensate the families of the authors or inventors, it was intended to protect the livelyhood of the author or inventor so he could go out and do more excellent work. It also allowed for the advancement of the state of the art by allowing others to expand on previous works.
Yeah, the Microsoft guy is a dufus, but he never said outlaw Open Source. I was anticipating reading that they were lobbying to prohibit Open Source software in the government but that wasn't the case. No wonder slashdot is home to trolls. The editors in chief are nothing more than trolls themselves.
Most of what he said was wrong or overstated or oversimplified, but Hemos' was no better.
It's not a flaw. It depends on what the desirable outcome is. If you're a business that sells software the desirable outcome is probably money. In this case your priority would most likely be developing software for Microsoft Windows. Ports to Linux, MacOS or BeOS might be planned and developed, but if those ports jeapordize the main project they'll at least be put on the back burner.
In Linus' case the desirable outcome is probably a stable kernel that meets the needs of most people who want to use it. In order to get to that stable kernel he has to mimimize outside influences that either distract him from the goal, or very likely compromise it. The amount of testing that needs to be done to ensure the integrity of a patch is proportional to the size of the patch. The ease of understanding a patch is inversely proportional. This is why he asks for small patches that do specific things as opposed to the tarball of patches that the Linux PPC folks try to force him to take. Any piece of that large patch may no longer be applicable in his development tree. A bunch of small patches may also not work, but at least they're isolated and much easier to debug.
I'm sure that eventually the PPC code will be merged in, but in the mean time given (from other posts including comments from Linus) that the modifications are being given in a format that he doesn't want he's doing the right thing. It's called project management.
It may be a case where fixing outstanding bugs is more important than maintaining the different architectures is the right thing to do for the time being. If you look through the kernel logs you'll see that from time to time architecture specific changes are rolled in. The number of x86 users dwarfs the number of PPC/SPARC/etc users, so in the time it takes to verify and integrate the other architectures might be better spent elsewhere for now.
I'd hope that architecture politics stay away from the linux kernel.
I think rebates as opposed to sales are done for accountancy purposes. The theory would be that these rebate offers drive up sales over a particular stretch of time. Of course they're discounting the item, cutting into their profit margin. The financial impact isn't realized until a later quarter, so they get a good short term increase in sales isolated from the decrease in profit margin. The wallstreet types like this.
I wouldn't be at all suprised if there weren't some tax benefits for the company either.
I'd rather have a closed source nVidia driver for performance reasons than an open sourced ones for political. The Matrox G### cards are pretty open, but the performance of them is only a fraction of the nVidia closed source drivers. Open Source is great, but isn't the solution to all problems. From appearances having fast OpenGL is one of those "isn't a solution" at least at this time.
Nope, satellite broadcast companies did this about 10 years ago (maybe more) to the big dish owners. The decryption boards had been hacked, they sent out a magic bullet which disabled them prior to some big event.
When VideoCypher was the technology used to protect content, which had been hacked so you could install keys for whatever channel you wanted, they also started changing the keys more often. First weekly, then daily, then several times a day. Since you had to manually enter those keys if you had a hacked board it helped improve their sales;)
You're receiving encrypted signals, you've got to go through an extra step to decrypt them. This should be legal (and probably isn't, but I don't write laws). In my opinion its also in the companies rights (and good judgement) to try to use technology to prevent this. That's what DirectTV did. They could've easily done a sting operation but they didn't. Wait till the dumb SOB's call in and complain they can't see the SuperBowl, give them a choice, a five year fully loaded DirectTV contract or a visit from the police. This has been done before.
Was it cost effective? Probably, most of the people with affected boxes probably weren't 'hackers', they just bought hacked boxes. A lot of them would gladly purchase a DirecTV to get their football fix.
I still think the DirectTV hack was beautiful, even if I would've been pissed off (sort of, but laughing) had I been one of the ones affected by it.
The number of people who can run out and fabricate an Open Source processor will hover at around zero. Does this mean that Open Sourcing a processor is somehow delusional?
No, not at all. Some corporations will make use of the processor core, possibly in interesting ways. Academic researchers will be have an already working design to build on, to expand, to explore new techniques. For large universities such as Stanford this won't be important, they have the resources to start a processor from scratch (and have, and have gone commercial with it, and did it again).
For research groups at smaller universities it can be a huge jump start. It's a working core, there's compilers for it already, there's simulators for it. They can do research in how the silicon architecture can be optimized for specific families of algorithms. Some of these algorithms will be fruitful, making some future generation of processor more efficient at some metric.
I've designed and fabricated a processor while doing my masters degree. Due to starting from scratch a very simple design had to be pursued. If we could've started from a working core we would've been able to do much more interesting work.
Let me guess. You think that somebody driving 5 miles per hour over the speed limit will suddenly spin out of control and burst into flames. You drive a safe 15 mph under the speed limit and shake your fist at all those impatient people who have to pass you. You drive with your head bent over your steering wheel since you're so nervous behind the wheel.
Please, before you kill somebody sell your car and buy a bus pass.
I will admit that speeding can be dangerous, but I'll also say that it can be safe. If you're in control of your vehicle you are a safe driver. If you're not in control of your vehicle you are unsafe, regardless of the speed. There are situations where driving below the speed limit is the right thing to do: icy roads, low visibility, heavy traffic. There are many times where driving above the speed limit is perfectly safe as well. It's all about knowing the limitations of your vehicle, the limitations of the people around you and your own limitations.
Sometimes the best way out of an otherwise dangerous situation is to speed. I've been in a situation a few times where a tractor trailer has decided they really really want to share my lane and moved in. If you're near the tail end of the trailer when this happens the smart thing to do is slow down. If you're near the front of the rig slowing down would be suicidal. The proper thing to do is step on the gas and get out of the way. The wrong, and quite possibly fatal, thing to do is to be in that lane when the truck completes its move.
Slashdot Mirror
Asynchronous logic uses a handshake signal to indicate completion of an operation. This is often done by adding an extra signal along with the result, the signal is asserted when the operation of the gate is completed.
There are other problems that have stopped asynchronous circuits from wide adoption. The handshake signalling adds overhead. More circuitry and more wiring, which means more silicon area used and higher wiring congestion.
A furnace is designed to be most efficient at radiating infrared energy. Other wavelengths are emitted but again they're just waste.
If you really believe that the amount of energy used by leaving on a lightbulb (or computer, or blender etc) continuously is offset by the furnace being operated on a lower duty cycle you're wrong. Just as wrong as if you insisted you don't need lights in the winter since the furnace is on and emitting visible light radiation.
If this ever becomes untrue I'll fully expect to live in an EZ-bake house.
The CAD tool part of it is easy to pick up isn't a fundamental part of the skills I consider important. When I'm interviewing candidates direct out of school my eyes glaze when they tell me how many software packages they know how to use or how many languages they've used. Tell me about a project you've been involved in, tell me how you went about the design project.
The tools I use now will be obsolete in a few years, sooner if I change companies. The skills I have will be valuable forever.
I consider knowing languages analogous to knowing how to flip burgers. Sure, you've flipped burgers at McDonald's, that doesn't mean you're a chef.
I think everything you need is in /etc/resolv.conf
nameserver n.n.n.n
nameserver m.m.m.m
nameserver the.unofficial.tld.people
The RIAA sold less CD singles last year, they also can't attribute that to Napster without going through the statistics to prove it or they're lying as well. Actually, they've already lied, Napster has lied and Napster supporters have lied.
Personally I'm pro mp3 sharing, moreso now that the RIAA has started bullying efforts. I'm not necessarily pro-Napster. It's a corporate entity trying to make money enabling other people to share music they have no copyright on. Get rid of the corporate pig in the middle and I fully support it. I think the record industry and electronic distribution can co-exist, but it will be a less powerful record industry.
If anybody can find inexpensive hardware MPEG2 encoders that have linux support I'd really appreciate it, as I've started playing with digital video and my DV camcorder.
The TiVO has hardware MPEG2 encoding and runs linux but I don't think you can get the drivers or hardware.
I've improved myself somewhat by following through on part of Betty Edward's The New Drawing on the Right Side of the Brain but I still don't have the innate ability. It's still pretty interesting though, the sense of everything disappearing from my consiousness except for the artwork is very similar to what happens for me when working on a difficult and interesting engineering problem. I'm going to finish the book when I'm less overwhelmed with work.
Computer Engineering is more about the how's and why's of the guts and external interfaces of a computer system. How do architectural decisions effect performance and scaling for instance? How do different cache sizes and mechanisms effect performance for given applications? You'll spend time writing software and you'll spend time designing hardware and often you'll be doing both at the same time.
You can avoid university and still do well, some notable people have taken this route. Most become code monkeys however, they implement others work rather than architect the systems (software and/or hardware) that they work on.
I'm an electrical engineer with a lot of computer science courses, so I'm a bit of an outsider. I've seen the curriculum of my more (and less) succesful friends though, and the above is my take on the succesful programs.
Every nVidia product is next release in a feature reduced version, often available in white box OEM versions, for a fraction of the cost. Just like the GeForce 2 MX there will be a GeForce 3 MX.
The performance won't be as good, but it'll be available for a more reasonable price.
I no longer visit a number of sites because they display pop-up ads, JAVA (I assume anyway, they're interactive) ads which crash or slow my browser, or sites that display more advertisement than material. I could set things up so that the annoyances wouldn't appear, but I'd rather just ignore the site. Oh yeah, add rapidly flashing ads to the list of things that will turn me off (i.e., the morons at X-10)
All of this can be distilled down to an observation that advertisers should listen to. No matter how slick you think the advertisement is, no matter how many hours you spent on it or no matter how "ground breaking" it is, you've failed miserably if you turn off your target audience.
Market droids should also realize that it isn't the size of their adds that are stopping me from clicking through, its the fact that I already know the product or service or I'm just not interested.
United States patent law includes provisions that put patented inventions in the public domain after a period of time has elapsed. This provides freedom for people to build on, extend and improve on the patented entity.
If you're going to play dictionary definition rebuttal then I'll propose that the GPL is by definition... not free.
The GPL restrains me from taking the linux kernel and marketing it as the closed source product eXtro's K-rad Kernel.I don't agree that I should be able to do this, but it is an example of your practice of using the dictionary as a bludgeon to quell observations you don't agree with. In this case it "proves" the untrue assertion that the GPL is anti-freedom.
What about coal miners? The point is that there are a lot of people doing dangerous jobs for not very much money. We may never know his motivation entirely, until they start to delve into his personal affairs. Gambling debts? Mistresses? Illegitimate children? Money probably wasn't the real factor, more likely it was a needed influx of money to bury some other problem.
Just conjecture, but from everything I've read recruiting for espionage (whether corporate or national) is usually rooted in finding some personal problem and exploiting it with the proper carrot.
If you're a newlywed and getting rid of your vast pornography collection a format is probably sufficient. Rewrite over the data with a random pattern if you're paranoid your new wife will use Norton Marriage-Breaker 2K1.
If you're dealing in contraban, whether the contraban is on the harddrive or the harddrive just contains incriminating evidence, destruction of the device would be your best bet. Remove the platters, stick them in the microwave for a few seconds (mostly just for the fireworks), sand them, sand blast them and melt them.
Just like cryptography the lengths you should go to depend on the difficulty in going to those lengths, the probability that those lengths can be compromised and the cost if those lengths are compromised.
With or without copywrite Grant still could've written his memoirs and expected them to provide for his family. They've got something, his memoirs, which others don't have. It's up to them (or him prehumously) to negotiate favourable terms.
The copyright law wasn't intended to protect or compensate the families of the authors or inventors, it was intended to protect the livelyhood of the author or inventor so he could go out and do more excellent work. It also allowed for the advancement of the state of the art by allowing others to expand on previous works.
Most of what he said was wrong or overstated or oversimplified, but Hemos' was no better.
In Linus' case the desirable outcome is probably a stable kernel that meets the needs of most people who want to use it. In order to get to that stable kernel he has to mimimize outside influences that either distract him from the goal, or very likely compromise it. The amount of testing that needs to be done to ensure the integrity of a patch is proportional to the size of the patch. The ease of understanding a patch is inversely proportional. This is why he asks for small patches that do specific things as opposed to the tarball of patches that the Linux PPC folks try to force him to take. Any piece of that large patch may no longer be applicable in his development tree. A bunch of small patches may also not work, but at least they're isolated and much easier to debug.
I'm sure that eventually the PPC code will be merged in, but in the mean time given (from other posts including comments from Linus) that the modifications are being given in a format that he doesn't want he's doing the right thing. It's called project management.
I'd hope that architecture politics stay away from the linux kernel.
I wouldn't be at all suprised if there weren't some tax benefits for the company either.
I'd rather have a closed source nVidia driver for performance reasons than an open sourced ones for political. The Matrox G### cards are pretty open, but the performance of them is only a fraction of the nVidia closed source drivers. Open Source is great, but isn't the solution to all problems. From appearances having fast OpenGL is one of those "isn't a solution" at least at this time.
When VideoCypher was the technology used to protect content, which had been hacked so you could install keys for whatever channel you wanted, they also started changing the keys more often. First weekly, then daily, then several times a day. Since you had to manually enter those keys if you had a hacked board it helped improve their sales ;)
You're receiving encrypted signals, you've got to go through an extra step to decrypt them. This should be legal (and probably isn't, but I don't write laws). In my opinion its also in the companies rights (and good judgement) to try to use technology to prevent this. That's what DirectTV did. They could've easily done a sting operation but they didn't. Wait till the dumb SOB's call in and complain they can't see the SuperBowl, give them a choice, a five year fully loaded DirectTV contract or a visit from the police. This has been done before.
Was it cost effective? Probably, most of the people with affected boxes probably weren't 'hackers', they just bought hacked boxes. A lot of them would gladly purchase a DirecTV to get their football fix.
I still think the DirectTV hack was beautiful, even if I would've been pissed off (sort of, but laughing) had I been one of the ones affected by it.
No, not at all. Some corporations will make use of the processor core, possibly in interesting ways. Academic researchers will be have an already working design to build on, to expand, to explore new techniques. For large universities such as Stanford this won't be important, they have the resources to start a processor from scratch (and have, and have gone commercial with it, and did it again).
For research groups at smaller universities it can be a huge jump start. It's a working core, there's compilers for it already, there's simulators for it. They can do research in how the silicon architecture can be optimized for specific families of algorithms. Some of these algorithms will be fruitful, making some future generation of processor more efficient at some metric.
I've designed and fabricated a processor while doing my masters degree. Due to starting from scratch a very simple design had to be pursued. If we could've started from a working core we would've been able to do much more interesting work.
Let me guess. You think that somebody driving 5 miles per hour over the speed limit will suddenly spin out of control and burst into flames. You drive a safe 15 mph under the speed limit and shake your fist at all those impatient people who have to pass you. You drive with your head bent over your steering wheel since you're so nervous behind the wheel.
Please, before you kill somebody sell your car and buy a bus pass.
I will admit that speeding can be dangerous, but I'll also say that it can be safe. If you're in control of your vehicle you are a safe driver. If you're not in control of your vehicle you are unsafe, regardless of the speed. There are situations where driving below the speed limit is the right thing to do: icy roads, low visibility, heavy traffic. There are many times where driving above the speed limit is perfectly safe as well. It's all about knowing the limitations of your vehicle, the limitations of the people around you and your own limitations.
Sometimes the best way out of an otherwise dangerous situation is to speed. I've been in a situation a few times where a tractor trailer has decided they really really want to share my lane and moved in. If you're near the tail end of the trailer when this happens the smart thing to do is slow down. If you're near the front of the rig slowing down would be suicidal. The proper thing to do is step on the gas and get out of the way. The wrong, and quite possibly fatal, thing to do is to be in that lane when the truck completes its move.